chapter 2 from meaning to form - forsiden - universitetet i … · chapter 2: from meaning to form...

Chapter 2: From meaning to form

1

Chapter 2

From meaning to form The figure below shows how one of the founding fathers of modern linguistics, Ferdinand de Saussure, envisaged the speech situation:

Person A produces speech sounds to transmit ideas from his mind to the mind of person B. Person B more or less successfully reconstructs the ideas of person A in his own mind on the basis of the sound waves that his ears pick up. Since most people are not mind-readers, they need a medium for communication, and language is such a medium.

The traditional branches of linguistics cover different stages on the way from ideas to sound waves (and back again). At one end, we have semantics, the study of linguistic meaning. At the other end, we have phonology, the study of speech sounds. From one perspective, meaning is what language is all about, the other branches of linguistics being concerned with how meaning is given form, so that it may be conveyed between speakers and hearers. Sounds, on the other hand, are not strictly necessary, and some languages, most notably the sign languages of the deaf, convey linguistic meaning without the use of speech sounds. In the vast majority of languages, however, sounds play a crucial role.

In between semantics and phonology, there are two other branches of linguistics, and both are concerned with the structural principles for the combination of meaningful elements. In traditional terminology, morphology is the study of word structure, of how smaller units are combined into words (as in English sing-er-s), while syntax is the study of sentence structure, of how words are combined into phrases (like the old man), clauses (like that he was dying) and sentences (like The old man knew that he was dying). Syntax and morphology have partly overlapping tasks, and languages differ radically in the degree to which they make use of these two branches to convey meaning. Some languages, such as Vietnamese, have little morphology and much syntax, whereas other languages, such as Inuit (Eskimo), have


2

much morphology and comparatively little syntax. In this book, the term grammar will be used as a cover term for morphology and syntax. 1

2.1 Lexicon and grammar All languages make use of a lexicon and a grammar. The lexicon is a mental dictionary containing all lexical items (such as words and fixed expressions) in a given language. The grammar is a set of rules for the usage of these lexical items, especially for ways of combining them with each other. 2.1.1 Lexical items A lexical item consists in a relation between meaning and form.2 The meaning of a lexical item is different from its reference. For instance, the English word tree represents a mental concept rendered by the Oxford English Dictionary as 'a perennial plant having a self-supporting woody main stem or trunk (which usually develops woody branches at some distance from the ground), and growing to a considerable height and size'. This concept is something that exists in the mind of a speaker of English; it is a psychological entity. The actual trees found in the physical world all belong to the reference of the word tree, not to its meaning, though there is obviously a close connection between the two. As a matter of convention, meanings are put between simple citation marks: 'tree'. A distinction between psychological and physical entities is also made in the case of form. Although the phonetic form of a lexical item is concrete and physical, consisting of sound waves produced and modulated by the various speech organs, the sound shape of a word also has its psychological aspects. For instance, although the sounds [th] (aspirated t, as in tea) and [t] (unaspirated t, as in steal) are phonetically different (just hold your hand in front of your mouth to feel the difference between the two), this difference is not enough to make speakers of English feel that they are different sounds. This contrasts with speakers of Chinese, for whom the same distinction is vital, because it constitutes the only difference between words like [thwó] 'pull' and [twó] 'many; much'.3 In Chinese, therefore, the distinction between [th] and [t] has a psychological reality that it lacks in English. A physical sound is called a phone and put between square brackets ([th] and [t]), while a psychological sound is called a phoneme and put between slashes (/th/ and /t/). We shall return to this distinction, and the problems it involves, later in this chapter. Basically, the relation between the meaning and the form of a lexical item is arbitrary. There is no good reason why a tree is called tree in English, except that most speakers of English agree that this is the case. It is a matter of convention. Other languages follow other conventions. On the one hand, the same meaning may be expressed through different forms, as when 'tree' is called arbre in French, mti in Swahili, ju in Japanese and tlugvi in Cherokee. On the other hand, one and the same form may express different meanings, as when the syllable written or transcribed ni is used to represent the number 'nine' in Norwegian and Danish, the dative particle

1 Some linguists prefer the term ’morphosyntax’ and use ’grammar’ in a wider sense, to refer to all the underlying rules of language, including semantics and phonology. 2 Saussure uses the terms ’sign’, ’signified’, and ’signifier’ (’signe’, ’signifié’ and ’signifiant’ in French) in basically the same meanings as ’lexical item’, ’meaning’ and ’form’. 3 Written tuo1 and duo1 in pinyin transcription.


3

(corresponding to English ‘at’, ‘on’, ‘in’, ‘to’ etc.) in Japanese, the first-person pronoun 'I' in Hausa and the second-person pronoun 'you' in Chinese. In all these cases, there is no obvious motivation for the relation between meaning and form. There are many interesting exceptions, and we will return to some of them in chapter 3, but it remains true that the relation between the meaning and the form of a lexical item is basically arbitrary and conventional. 2.1.2 Grammatical structure Language does not only consist of individual lexical items, but also of rules for the usage and combination of these items. Such rules also have both meaning and form. Consider, for instance, the following two sentences: A. Peter will come. B. Will Peter come? Both sentences are combinations of the same words, but they differ both in meaning (statement vs. question) and form (word order).4 In this case, the meaning is expressed through the structure of the sentence, not directly through its sounds. Of course, the structure also has consequences for the phonetic shape of the sentence, but the crucial difference does not lie in the phonetic difference between p-e-t-e-r-w-i-l-l and w-i-l-l-p-e-t-e-r, but between the order of the subject and the following auxiliary.5 Compare the following two sentences: C. Mary must die. D. Must Mary die? With regard to meaning, the difference between C and D is exactly the same as the difference between A and B. The same is true about the difference in structure, with both pairs representing a contrast between subject-auxiliary-verb and auxiliary-subject-verb word order. With regard to phonetic shape, however, the difference between C and D (m-a-r-y-m-u-s-t vs. m-u-s-t-m-a-r-y) is not the same as that between A and B. There are several differences between lexicon and grammar. The lexicon contains particular items that must be remembered individually (such as Peter, will and come), while the grammar contains general rules (such as the rule stating that a question may be formed by moving the auxiliary in front of the subject). Lexical form consists in strings of sounds (such as /:pi˘t´/, /wIl/ and /k√m/), while grammatical form also involves structural patterns (such as the ordering of subject and auxiliary). Also, the lexicon tends to treat each item as a whole, while the grammar is always concerned with a combination of meaningful elements. Objects Form Complexity Lexicon individual items strings of sounds whole Grammar general rules structural patterns combination 4 For the sake of simplicity, we shall look away from the fact that they also differ in intonation pattern, or, in writing, in the use of punctuation marks. 5 The terms ’subject’ and ’auxiliary’ are discussed later in this chapter.


4

The distinction is not absolute. First, what is a rule? Is the auxiliary will a lexical item meaning 'to be going to' or a grammatical marker employed by the rule for future marking in English? Second, grammatical meaning is expressed not only through structure, but also through affixes (like plural -s) and function words (like the perfect marker have). And third, the lexicon also contains a large number of complex items, such as irregularly inflected words like children, derived words like kindness, compounds like milk-shake or idioms like kick the bucket. In such cases, grammatical structure also enters into the lexicon. In fact, information about the grammatical properties of each lexical item, such as word class (noun, verb, adjective etc.), is an important part of the lexicon. The fact that there is some overlap, however, does not make the distinction between lexicon and grammar less important. As we have seen, the connection between meaning and form in simple lexical items is usually arbitrary. This is much less true of grammatical structure, which is often at least partly motivated. While both milk and shake are arbitrary names, the compound milk-shake is not, since knowing the words milk and shake is enough to indicate, albeit imprecisely, the meaning of the word milk-shake. And while the plural form men is arbitrary, the plural form hens is not, since knowing the word hen and the plural affix -s is enough to determine the meaning of hens. ?? Furthermore, some grammatical structures are not only motivated, but iconic in the sense that they function as images of the reality that they refer to. The most obvious case is temporal iconicity. In the following famous sentence, reportedly uttered by Julius Caesar after he had conquered the Pharnaces, the sequence of the three clauses reflects the temporal sequence of the events referred to: I came, I saw, I conquered. If the sequence of the clauses were changed, so would the temporal sequence of the events referred to. The temporal iconicity principle is extended to cover a number of other grammatical structures: 1. Given information tends to precede new information (since in the mind given information exists prior to new information). 2. In conditional sentences, the if-clause tends to precede the then-clause (since the if-clause usually refers to events that exist prior to the events of the then-clause). 3. Clauses expressing cause tend to precede clauses expressing effect (since a cause precedes its effect in time). None of these principles are unbreakable, but they seem to apply to a greater or lesser extent in all known languages across the world. 2.1.3 Concepts We noted above that meanings are mental concepts. The meaning of the word ball, for instance, is a mental idea of what it takes to be a ball, defined by the Oxford English Dictionary as a ’globular body’. Of course, such concepts are assumed to have a reference, to refer to some kind of external reality, either in the physical world or in other mental or even spiritual realms. (What is, for instance, the reference of the word God?) Meaning and reference, however, are not the same. The meaning of the


5

Is a non-globular ball a ball?

word ball is not the actual balls that happen to roll around (or fail to do so) in our physical world (or even in our fantasy or in some spiritual realm), but, again, our concept or idea of what it implies to be a ball, to be a globular body. Many concepts are clearly socially constructed, formed by the child (or, sometimes, the adult) in the process of learning how to relate words and grammar to reality. To form a concept of a computer, for instance, one needs a certain experience or, at least, extensive explanation. In the case of ball, however, some linguists would argue that the concept of a globular body is not, or at least not only, socially constructed, but innate. While most of us have hardly ever seen a perfectly globular body, all of us have a concept of what it means for something to be one. Is that because we are born with this idea? Some would say yes and insist that the basic elements of semantics are universal and innate. Others would say no and insist that even if the concept of a ball should turn out to be universal, that is not because of some innate idea, but because the experience of (nearly) globular bodies is common to mankind all over the world. Whether or not some basic concepts are innate, however, the complex webs of meaning in natural languages are undoubtedly socially constructed. An innate idea of a globular body will not help a speaker of English to understand how the word ball can be used to refer to an American football or an English rugby ball, none of which is particularly globular.

All concepts belong to one or more semantic domains, to larger areas of linguistic meaning. For instance, seconds, minutes, days, weeks, and years belong to the the temporal domain; red, green, blue, and yellow belong to the colour domain, offside and penalty kicks belong to the football domain, and ski wax and pistes belong to the skiing domain. Some domains, like the temporal domain and the colour domain, are universal, while others, like the football domain and the skiing domain, are only found in certain cultures. Domains are of many different kinds; some are quite general, while others are highly specific.

Dictionary definitions are often based on such domains. In the Oxford English Dictionary, the meaning of the noun milk is explained as follows: The meaning of milk A whitish fluid, rich in fat and protein, secreted by the mammary glands of female mammals (including humans) for the nourishment of their young, and taken from cows, sheep, etc., as an article of the human diet. Among the domains that this definition is based upon are those of color, anatomy, zoology, alimentation, and dairying. Without some basic knowledge about these domains, we cannot understand all the aspects of this definition. Of course, not all speakers of English have such a broad understanding of what milk is. As children of 4 or 5 years we may only know that milk is a white liquid that is kept in the refrigerator and that our parents give us to drink in a cup. Another way of studying concepts is by analyzing them into smaller components, often referred to as semantic features (or components or properties).


6

Such analysis builds on the similarities and differences between various concepts. For instance, the concepts of ‘man’, ‘woman’, ‘boy’ and ‘girl’ all share the feature HUMAN, ‘man’ and ‘woman’ share the feature ADULT, ‘boy’ and ‘girl’ share the feature YOUNG, ‘man’ and ‘boy’ share the feature MALE, and ‘woman’ and ‘girl’ share the feature FEMALE:

MALE FEMALE ADULT ‘man’ ‘woman’

HUMAN YOUNG ‘boy’ ‘girl’

The features MALE and FEMALE clearly belong within the same axis, the semantic category of gender, while the features ADULT and YOUNG belong within another axis, the semantic category of age. Such analysis into semantic categories and features is useful in bringing out both parallels and differences between sets of concepts. For instance, there are both clear parallels and one clear difference between the concepts for human beings above and the terms for ovine, bovine and equine beings below:

MALE FEMALE ADULT ‘ram’ ‘ewe’

OVINE YOUNG ‘lamb’

MALE FEMALE

ADULT ‘bull’ ‘cow’ BOVINE

YOUNG ‘calf’

MALE FEMALE ADULT ‘stallion’ ‘mare’

EQUINE YOUNG ‘foal’

All four sets of concepts are analyzable in terms of the same semantic categories and features: gender (MALE vs. FEMALE) and age (ADULT vs. YOUNG). The difference is that young humans are subdivided into male and female, wheres the young non-human animals are not. In other words, the concepts for young human beings are marked (or specified) for gender, whereas the concepts young ovine, bovine and equine beings are unmarked (or neutral or unspecified) for gender.6 Strictly speaking, we should distinguish more clearly between concepts and the actual words of a living language. Although the English language does not distinguish between terms for ‘male lamb’ and ‘female lamb’, ‘male calf’ and ‘female calf’, ‘male foal’ and ‘female foal’, this does not necessarily mean that speakers of English do not have these mental concepts. Most speakers of English are certainly

6 Note some other differences between the four sets of terms analyzed here: In humans, the terms marked for gender and age are all commonly used in everyday language, while the unmarked term person is slightly more technical. In ovines and equines, the unmarked terms sheep and horse as well as the young terms lamb and foal are common, while ram and ewe as well as stallion and mare are more specialized. In bovines, there is no unmarked term at all, the species usually being referred to by the adult female term cow, though ox and calf are also common terms. Cf. the discussion of basic level terms later in this chapter.


7

able to form such concepts. It is only that their vocabulary does not prompt them to do so. From what has been said above, one might get the impression that each semantic domain contains a number of concepts, each of which may in turn be analyzed into semantic features: In fact, however, semantic domains (like color, anatomy, zoology, alimentation, and dairying) are also concepts, and so are semantic features (like MALE, FEMALE, ADULT, and YOUNG). To a large extent, semantic analysis implies analyzing concepts in terms of other concepts. Furthermore, the difference between semantic domains and semantic features should not be exaggerated. In fact, features like HUMAN, OVINE, BOVINE, EQUINE, MALE, FEMALE, ADULT, and YOUNG may also be used to form classes of concepts that share the same feature, so-called semantic classes. These are very similar to what we have called semantic domains. 2.1.4 Conceptualization In addition to being shaped by our social experience and our genetic endowment, our concepts are also partly shaped and reinforced by our language. Different languages conceptualize the same reality in different ways. English, for instance, treats blue and green as different colours, while other languages, like the Mexican language Tarahumara, treat them as variants of the same colour. English has one term for ’cousin’, while many European languages distinguish between male and female cousins (French cousin vs. cousine, Spanish primo vs. prima, German Vetter vs. Base), and Chinese has eight different terms for ’cousin’ based on gender (male vs. female), age (older vs. younger) and whether or not they have the same family name (paternal vs. maternal). Learning a new language, therefore, often implies learning new conceptual distinctions, like the one between ’blue’ and ’green’, ’male cousin’ and ’female cousin’, ’paternal cousin’ vs. ’maternal cousin’, etc. In other cases, we learn to look away from conceptual distinctions routinely made by our own language. Speakers of French, Spanish and German who learn the English word cousin, also learn a new concept, viz. one that is unmarked with regard to gender. Chinese-speaking learners of English will have to learn to look away from not only the distinction between male and female, but also that between older and younger, as well as paternal and maternal cousins, since Chinese also lacks an unmarked term corresponding to English cousin. And while English makes a conceptual distinction between ’blue’ and ’green’, it lacks an unmarked term corresponding to Tarahumara siyóname (which covers both blue and green).

semantic domain

concept1 concept4

concept3

concept2

semantic feature1 semantic feature4 semantic feature3 semantic feature2


8

Basically, language contributes to conceptualization in two ways, through lexicon and grammar. A concept is given lexical expression (or it is being lexicalized) when a word (or some other lexical item) is used to represent it, as when German Vetter and Base lexicalizes the concepts of ‘male cousin’ and ‘female cousin’. A concept is given grammatical expression (or it is being grammaticalized or grammaticized)7 when a morphological or syntactic construction is used to represent it, as when English marks the conceptual distinction between ‘substance’ and ‘object’ by distinguishing between mass nouns (water, clay, love) and count nouns (house, sculpture, idea). Morphologically, mass nouns have no plural form, while count nouns do; syntactically, mass nouns cannot be directly preceded by numerals or words like many, while count nouns can. A large number of languages across the world do not make an equally clear distinction between the two types, because their nouns do not have plural forms and may not be directly preceded by numerals, and because the same word is used for much and many. We will come back to the distinction between mass nouns and count nouns later in this chapter. Whether or not a concept is lexicalized in a given language is not always an either-or question. It makes sense to talk of degrees of lexicalization. For instance, it has been proposed that all languages lexicalize the concepts of 'man' and 'woman'. In most languages this is done by means of simple words like English man and woman.8 The corresponding Chinese words, however, are complex terms consisting of the word nán 'masculine' or nü# 'feminine' plus the word rén 'person':

na2n-re2n 'man' nü#-re2n 'woman'

Japanese goes one step further and adds the grammatical particle no between otoko 'masculine' or onna 'feminine' and hito 'person', marking clearly that each expression consists of two separate words (in addition to the particle), the first of which modifies the second: otoko no hito 'man' onna no hito 'woman' Even in the Japanese case, however, one may still argue that otoko no hito and onna no hito are fixed expressions, and that the concepts of 'man' and 'woman' are lexicalized, although they are represented by fixed, idiomatic phrases rather than single words. The degree of lexicalization, however, is much weaker than in English. We get the following correspondence between linguistic form and degree of lexicalization:

7 The terms ’grammaticalization’ and ’grammaticization’ are, however, more often used in another meaning, viz., the historical development of a word into a grammatical marker, as when the English numeral corresponding to modern one developed into the article an or a. 8 Historically, even woman is a complex term, going back to Old English wïfman, from wïf ’woman’ and man ’person’.


9

Degrees of lexicalization

STRONG

MEDIUM

WEAK

Linguistic form

simple word

complex word

fixed phrase

Examples

man

woman

nán-rén nü#-rén

otoko no hito onna no hito

The difference in degree of lexicalization is reflected in actual language use. While the English terms man and woman are often used even if the gender of the person involved is irrelevant, the Chinese terms na2n-re2n / nü#-re2n and the Japanese terms otoko no hito / onna no hito are normally only used when the gender of the person is highlighted. In contrast to its Chinese and Japanese equivalents, the English word person sounds slightly formal and technical. Thus, although it may be true that the concepts of 'man' and 'woman' are lexicalized in all living languages, they are so to different degrees and in different ways. 2.2 Meaning One way of studying meaning is to do it on the basis of the relationship between language and the world surrounding us. As learners of foreign languages, we often find ourselves in the following two situations: THE SEMASIOLOGICAL SITUATION THE ONOMASIOLOGICAL SITUATION We present a word and ask a native speaker, “What kinds of objects can it appropriately be used about?”

We present an object and ask a native speaker, “What words can appropriately describe it?”

In the following, we shall start by looking at meaning from these two perspectives. 2.2.1 Semasiology and prototypes In the semasiological situation, our perspective goes from language to the world. Let us imagine that you are talking to Li, a native speaker of Chinese, in his home. You ask him to show you an appropriate instance of the kind of object that is designated by the Chinese word be\izi ’drinking vessel’. Hopefully, Li will show you a drinking vessel of some kind, but there are many kinds of drinking vessels. They vary in material, size, and shape, they come with and without a handle or two, with and without a lid etc. Some of the variation is shown in PICT. 1.


10

PICTURE 1. Drinking vessels Probably, Li will be of the opinion that some drinking vessels are better representatives of the category that be\izi is used about than others, and, since he wants to help you in your study of the Chinese language, he will show you a good representative of the category, an instance that we may talk about as a prototypical instance. Most concepts have prototypical instances, or prototypes, entities to which a word is typically applied. We may guess that a be\izi made of porcelain (also known as china!) will be closer to Li’s idea of the prototypical be\izi than those made of glass, paper, silver or some other material. He will probably point to something resembling a cup, like one of those in the upper row in PICTURE 1, especially the two in the middle. In fact, be\izi is translated into English as ‘cup’ in most Chinese–English dictionaries, but this translation is based on the prototypical meaning, and in fact all the drinking vessels in PICTURE 1 can be referred to as be\izi. It is tempting to translate be\izi into English as ‘drinking vessel, and in particular a porcelain cup’ as a reflection of its usage. Just as concepts may differ from language to language (be\izi covering both English cup and glass), so may prototypical representatives of the same concept. In one study, for instance, it turned out that native English speakers in Britain tended to regard the potato as a protypical vegetable, while many immigrants in the country did not regard it as a proper vegetable at all. 2.2.2 Onomasiology, hyponymy and basic level terms In the onomasiological situation our perspective goes from the world to language. You show Li a cup, and ask him to tell you what Chinese word(s) he would regard as an appropriate name or designation for it. We expect him to mention at least the word be\izi. But there are many other possible designations. Li might have called it cha;be\i ‘teacup’, qìmûn ‘vessel’, or even do\ngxi ‘thing, object’! The four words cha;be\i, be\izi, qìmûn, and do\ngxi may all be used correctly about the cup. These words belong to different levels in a system of classification, with an increasing degree of generality:


11

teacups, cups, and vessels are kinds of things; teacups and cups are kinds of vessels, and teacups are kinds of cups. In semantics, a less general term is a hyponym of the more general term, and this phenomenon is referred to as hyponymy. Jiu'be\i ‘winecup’ and cha;be\i ‘teacup’ are hyponyms of be\izi ‘drinking vessel (and in particular a china cup)’, which in turn is a hyponyms of qìmûn ‘vessel’. The more general term is referred to as a hyperonym or simply a superordinate term, so that qìmûn ‘is superordinate to be\izi, and be\izi is superordinate to jiu'be\i and cha;be\i. Li did not refer to the cup as be\izi just by coincidence, but because cups are habitually called be\izi in Chinese. In the hierarchy of terms that can be applied to cups, be\izi is the basic level term, the term used unless there are good reasons to do otherwise. The basic level is the highest level at which we can form a mental image of a concept. We can easily form a mental image of a teacup or a cup, but vessels vary too much in shape, and we have to go back down to the basic level. Terms above the basic level tend to have a technical or scientific flavor, which is the case with Chinese qìmûn ‘vessel’.

Let us imagine that you are asking Aamadu, a native speaker of Fula, what he sees on PICT. 2. He will probably tell you that it is a nagge ‘cow’. But he could also have called it jamale ‘cow with big patches of black and white’ or wumale ‘cow without horns’. Alternatively, he could have stated that it is a picture of ndabbawa wuro ‘domestic animal’ (literally, ‘animal of the village’), simply ndabbawa ‘animal’, or huunde, a noun meaning ‘animal’ or ‘thing’ more generally.9

But nagge is clearly the basic level term applied to the «entity» in PICT. 2 (just as cow is the basic level term in English). It is the term a native speaker of Fula comes up with when seeing the picture. The terms jamale and wumale are technical terms for cow experts (who, admittedly, there are lots of among the Fula, traditionally a people of cattle nomads), while simply telling that it is ndabbawa wuro ‘domestic animal’ or ndabbawa ‘animal’ sounds like a conscious attempt to be less informative than expected. We can easily form a mental image—or make a drawing—of a cow, while the same cannot be said of a domestic animal, which is a generalization across cows, horses, sheep and goats. These latter animals, on the other hand, are also easy to conceptualize.

Basic level terms vary from language to language. For instance, both English cow and Fula nagge are gender-specific terms, while the corresponding basic level term in Chinese, niú, is not specified with regard to gender, and its hyponyms go1ng-niu2 ’ox’ and mu3-niu2 ’cow’ are clearly less basic. As it happens, it is a general tendency for Chinese basic level terms to be gender-neutral where English and many other languages have gender-specific terms. As we saw above, man and woman are basic level terms in English, the hyperonym person being slightly more technical in flavour, while the Chinese basic level term is the gender-neutral rén ’person’, the hyponyms nán-rén ’man’ and nü#-rén ’woman’ being less basic and therefore much less commonly used.

9 Huunde means ‘animal’ explicitly in expressions like huunde ladde ’wild animal’ (literally, ‘animal/thing of the bush’).

PICTURE 2


12

2.2.3 Polysemy, homonymy, and synonymy Meaning is not only about relations between language and the outside world, but also between elements within the language itself. We have already seen this in the case of hyponymy discussed above. In the present subsection, we shall look at three other types of semantic relations. The first of these is polysemy, the phenomenon of a single word having several related meanings. Take, for instance, the English word date, which may refer to, among other things, a point in time or an appointment (especially an appointment with romantic undertones). The relationship between ’point in time’ and ’appointment [at a certain point in time]’ is obvious; these meanings are clearly related. In these senses, therefore, date is only one polysemous word:

FIGURE 2. The polysemy of the English noun date: one pronunciation, two (or more) related meanings

The same word may even be used to refer to the person with whom one has an appointment, especially if the appointment is a romantic one. The relationship between ’appointment’ and ’person with whom one has an appointment’ is also clear, and again we have a case of polysemy. Polysemy may be fruitfully contrasted with homonymy, the phenomenon of two words with unrelated meanings having, by accident, the same pronunciation. As it happens, the English spelling date (and its pronunciation [dEit]) may also represent another word, which refers to the fruit that grows on a date palm. The meaning of this word is clearly unrelated to the meanings of the word date ’point in time; appointment; person with whom one has an appointment’. The two words date are homonyms:

word 1: date ‘point in time; appointment; person with who one has an appointment’ same pronunciation different, unrelated meanings

word 2: date ‘fruit from the date palm’ FIG. 3. The homonymy of the two English words date: one pronunciation, two unrelated meanings

Both words are in fact polysemous, the fruit term also sometimes referring to the whole fruit tree (the date palm). Homonymy, therefore, holds between independent words each of which may be polysemous:

HOMONYMY word 1: date word 2: date

’point in time’ ’appointment’ ’date fruit’ ’date tree’ POLYSEMY POLYSEMY FIG. 4. The homonymy and polysemy of the two English words date

Actually, homonymy is not a semantic relation at all. The meanings in question are unrelated, though they happen to be expressed through the same form. When different words (with different pronunciations) have identical meanings, we have synonymy. The words date (in one of its meanings) and rendezvous are synonyms:

‘point in time’ one word: date related meanings

’appointment’


13

word 1: date ‘appointment’

different pronunciation same meaning word 2: rendezvous ‘appointment’

FIGURE 5. The synonymy between English date and rendezvous: two pronunciations, one meaning However, date and rendezvous are not totally synonymous, only partially so. Not only do they have different connotations, a date being much more casual than a rendezvous, but the two words do not have the same set of related meanings. The word rendezvous does not have the meaning ’point in time’, but it does have other meanings that date lacks, such as ’an appointed place of meeting’ and ’a base for naval ships or for military units’. Both words are polysemous, but in different ways: ‘point in time’

word 1: date related meanings ‘appointment’

different pronunciation same meaning ‘appointment’

word 2: appointment related meanings ‘base for naval ships or military units’

FIG. 6. The partial synonymy between English date and appointment In natural languages, total synonymy is very rare, if it exists at all. Synonymy nearly always means partial synonymy.

2.2.4 Metaphor and metonymy The English noun foot is highly polysemous, and has at least the following five meanings: The meanings of foot I. ‘the part of the leg beneath the ankle’ II. ’a length unit; 12 inches’ III. ‘a rhythm unit in poetic meter; a stretch of syllables of which one has primary stress’ IV. ‘lower part of an object (such as a mountain)’ V. ‘the part of a stocking that covers the foot’

On the basis of foot, we shall take a closer look at how semantic relatedness can be described in a more precise way. The meaning ‘the part of the leg beneath the ankle’ is placed as number I on the list. This is the primary (or basic or central) meaning, from which the other extended meanings are most naturally derived. There is an associative relationship between the primary meaning and meaning number II, since the length unit foot is originally based on the length of the body part. There is also an associative relationship between the primary meaning and meaning number III, since the rhythm unit foot reflects the common habit of using the foot to beat the rhythm while reciting poetry. Both of these relationships are instances of metonymy: Metonymy: an expression that basically designates one entity comes to be used of another entity in the same domain. The two entities are closely associated in space, time or cause, or by one being a part of the other.


14

Cases where one entity is part of the other are common. In now obsolete usage among anatomists, foot could refer to the whole leg. This is referred to as pars-pro-toto. The relationship between the primary meaning and meaning IV is different. The foot of a person and the foot of a mountain are not associated in space, time or cause, or by one being a part of the other. Instead, this usage of foot reflects the fact that we may, to some extent, structure a mountain in terms of our own bodies. Just as human feet are (usually) the lowest parts of the human body, so the foot of a mountain is the lowest part of a mountain. This is called metaphor: Metaphor: one domain is thought of or understood in terms of another domain. The human body is one domain and the topography of mountains is another. When mountains are thought of in terms of the human body, it follows that the lowest part of a mountain may be called its foot. The foot of a human being and the foot of a mountain are analogous entities in two different domains. Metaphors are mappings between different domains, where the structure of one domain, the source domain (like the human body) is imposed on another domain, the target domain (like the topography of mountains). As a general rule, the source domain is a more familiar knowledge structure than the target domain.

An interesting example of structure imposition from a familiar source domain to a less familiar target domain is found in the Fula dialect of Maasina in Mali, where the bicycle is called puccel njamndi, literally ‘little iron horse’ (from puccel ‘little horse’ and njamndi ‘iron’). This name indicates that the speakers have imposed a «horse structure» on the bicycle. On the basis of this creative and fruitful construal of a new technical device, the Fula speakers started using equine terminology for the different parts of the bicycle, as illustrated in TABLE 1. TABLE 1. Maasina Fula horse and bicycle terms TERMS MEANING IN THE SOURCE DOMAIN MEANING IN THE TARGET DOMAIN koyngal ‘foot, leg’ ‘wheel’ yitere ‘eye’ ‘lamp’ kirke ‘saddle’ ‘seat’

HORSE

BICYCLE

Let us finally return to the English word foot and compare its primary meaning to meaning V. On the one hand, the part of a stocking that covers the human foot is spatially associated with the human foot; thus we may see this as an instance of metonymy. On the other hand, as in the foot of a mountain, this part of the stocking is also its lowest part, and the two entities belong to different domains (body part vs. clothing); thus we may also see this as an instance of metaphor. 2.2.5 Sentence meaning Words are assembled into sentences. The meaning of a sentence is determined by the meaning of the words (or other lexical units) and grammatical structures of which it is composed.


15

Sentences designate situations, that is, mental representations of a state of affairs. To get an understanding of how language describes situations, the following four elements must be distinguished: THE FOUR ELEMENTS OF A SITUATION • The event • The participants in the event • The setting where the event and its participants are situated • The ground - the situation in which the sentence is uttered Sentence semantics typically concerns the associations between the elements of a situation. The event is linked to the participants. The event and the participants are linked to a setting. The event, the participants, and the setting are linked to the ground. EVENT Events are usually referred to with verbs. PARTICIPANTS The participants are entities playing different roles in the event, referred to as thematic roles (sometimes called semantic roles). The study of entities belongs to lexical semantics, while the roles entities play in an event is an important part of grammatical semantics. Roles differ somewhat from one verb to an other, so that the roles of beat are the beater and the beaten, and the roles of devour are the devourer and the devoured, but linguists have tried to make some generalizations. Some important roles are agent, instrument, patient, and benefactive.

The meaning of the prototypical verb as an energetic interaction where energy from an energy source (agent) is transferred to an energy sink (patient), as illustrated in the Arabic sentence in (9), meaning ‘the father [=agent] beat his son [patient]’.

The prototypical agent is an animate entity that instigates an active event with will and intention, while the prototypical patient is an inanimate entity that undergoes, is changed by, or is affected by the active event. A prototypical instrument is an inanimate entity by means of which an active event is carried out. It is used by the agent to affect the patient, and energy may be said to “flow” through it on its way from the agent to the patient, as depicted in FIG. 9. AGENT INSTRUMENT PATIENT energy source energy sink FIGURE 9. The energy flow from agent through an instrument to a patient. These three roles are illustrated by sentence (11), where ‘he’ is the agent, ‘thief’ is the patient, and ‘truncheon’ is the instrument: (11) He beat the thief with a truncheon.’ AGENT PATIENT INSTRUMENT


16

Furthermore, the benefactive is a participant who benefits or suffers from an event, as illustrated in sentence (12): (12) I slaughtered that ox for Buuba. AGENT PATIENT BENEFACTIVE SETTING Setting is a common denominator for the time, location, manner, reason, and purpose that an event and its participants are embedded in. In sentence (13), the adverb yesterday specifies the temporal setting: (13) I shot the fox yesterday with a gun.

TIME Location is illustrated in sentence (14): (14) I saw the girls on the bus. LOCATION

Manner is illustrated in sentence (15): (15) I walked slowly towards her. MANNER Reason and purpose are expressed in sentence (16): (16) Mary went to town to buy a circus ticket because she loves clowns. PURPOSE REASON GROUND What is talked about is anchored to the ground, which has the following elements: GROUND the speech event the participants of the speech event: the speaker and the addressee(s) the setting of the speech event. Most languages have words for the main components of the ground. In English, the speaker is I, the addressee(s) is/are you, the local setting of the ground is here, and the temporal setting of the ground is now. Let us take a look at the English sentence in (17). (17) I shot the fox yesterday with a gun. We can analyze the situation described in (17) in the following way: The event is expressed by shot. There are three participants, the agent (I), the patient (the fox), and the instrument (the gun). The temporal part of the setting is expressed by yesterday.


17

The event, the participants, and the setting are anchored to the ground in several ways:

The past tense form shot places the event at a point in time preceding the ground. The agent is identified as one of the participants in the ground, that is, the speaker: I. Definite articles give the information that the patient and the instrument are already known to the participants in the ground: the fox, the gun.

2.3 Words All languages have words, and most of us feel that we know what a word is. A wide variety of modern written languages mark word boundaries by means of spacing. But even English is not consistent in this respect. Should one, for instance, write tax payer, tax-payer or taxpayer? Why, according to the Oxford English Dictionary, is road map written as two words, while roadwork is written as one, and road-horse with a hyphen? And why are today and tonight written in one word, while to hand and to work (meaning ’at work’ in colloquial American English) are written in two words? Such idiosyncracies of English orthography prompt us to ask: What is a word? One possible answer is that a word is a meaningful unit with a relatively high degree of autonomy. Thus, happier is one word and more happy two, because -er is less autonomous than more, although they have the same meaning. In other words, the degree of cohesion between different elements within a word is much higher than that between different words within a phrase. The degree of autonomy of the word, as well as the degree of cohesion within the word, is reflected both in its grammatical and phonological properties. While happier is inseparable (nothing may be inserted between happi- and -er), more happy is not, as is clear from sentences like After my divorce, I have become more, not less happy. And while redneck has one main stress, red neck has two. Exactly how word boundaries affect grammar and phonology varies from language to language. The term ’word’ is useful in describing most languages in the world. The famous linguist Edward Sapir taught a speaker of the American Indian language Nootka to write his own language. The Indian had no difficulty in determining the words, and in the hundreds of pages of Nootka text that he wrote down, word boundaries are virtually always placed exactly where a linguist would place them. The centrality of the word, however, varies a lot across the world. Many languages do not have a non-technical word for ’word’. Chinese is one of them, and written Chinese does not mark word boundaries. Even when the language is written with the Roman alphabet, it is much more common to insert a space between every syllable than between every word. There is much disagreement even among specialists on how to parse a Chinese sentence into words. 2.3.1 Word classes Words belong to different word classes (or parts of speech), each class being characterized by a mixture of semantic and grammatical properties. Different languages have different word classes, though some word classes seem to be more or less universal.


18

NOUNS AND VERBS The two most basic word classes are nouns (like man, dog, sunflower, stone, water, clay etc.) and verbs (like kill, fall, grow, like etc.). The distinction between nouns and verbs is found in most languages. A few languages have been reported to lack this distinction, though this is controversial. In the following word pairs from the native American language Nootka, the same stem is used for nounlike and verblike meanings: [qo˘/as] ‘a man’ vs. [qo˘/as-ma] ‘[he] is a man’ [wAl•A˘k] ‘a going; a trip’ vs. [wAl•A˘k-mA] ‘[he] goes’. It remains true, however, that the vast majority of languages in the world do make a distinction between nouns and verbs. The basic semantic properties of nouns and verbs are more or less the same everywhere. Nouns typically refer to entities (objects or substances), while verbs typically refer to processes (events or states). Both entities and processes may be concrete, such as the entity ball and the process walk, or abstract, such as the entity democracy and the process become. Furthermore, processes may be entified (seen as entities) and referred to by nouns, e.g. singing (in his singing never ended). The grammatical characteristics of nouns and verbs vary a lot from language to language. In many languages, including English, nouns typically function as subjects and objects. English verbs may also be used as subjects or objects, but only in the infinitive (to sing) or the gerund (singing) forms. In English, nouns, but not verbs, may occur directly after the articles the and a(n). And while English verbs are inflected for tense, mood, person and number, nouns are only inflected for number, if at all. COMMON NOUNS VS. PROPER NOUNS One group of words that is usually included among nouns is names like Tom, Brazil, January and Adidas. While ordinary nouns are referred to as common nouns, such names are referred to as proper nouns. Like common nouns, proper nouns refer to entities, though they do not have a general meaning, only a specific reference. Like common nouns, proper nouns typically function as subjects and objects. But unlike common nouns, they do not occur directly after the articles the and a(n), and they are never inflected for number. The term ’noun’ is primarily used to refer to common nouns. COUNT NOUNS VS. MASS NOUNS In many languages, there is a clear distinction between count nouns and mass nouns. English count nouns, such as horse, share the following grammatical characteristics: - they have a singular and a plural form (horse vs. horses) - they may occur with words such as many or few (many/few horses) - they may occur with numerals (two horses) - they may occur with the indefinite article (a horse) In contrast, English mass nouns, such as smoke, share the following characteristics:


19

- they do not have a plural form (*smokes) - they occur with much or little rather than many or few (much/little smoke) - they do not occur with numerals (*two smokes) - they do not occur with the indefinite article (*a smoke) Count nouns typically refer to objects, while mass nouns typically refer to substances. • An object is a sharply delineated entity, and if it is divided into smaller pieces, these pieces are no longer instances of the same object. • A substance is not sharply delineated, and if it is divided into smaller pieces, these pieces are instances of the same substance. While a horse cut in two does not give two horses, smoke remains smoke even if divided in two. The distinction between count nouns and mass nouns plays a central role in the grammar of English, as it does in Swahili and many other languages of the world.10 There are also many languages, however, where it plays at best a peripheral role, since their nouns make no distinction between singular and plural forms, they make no distinction between many/few and much/little, numerals never precede nouns directly, and they have no indefinite article. This is true of a large number of languages along the Pacific Rim, in East and Southeast Asia, Polynesia and native America. CLASSIFIERS In languages where the distinction between count noun and mass noun plays a peripheral role, numerals do not directly modify any nouns. In these languages, numerals are typically followed by a classifier (sometimes called a measure word). Classifiers resemble English nouns like cup, piece and spray in a cup of tea, a piece of meat and a spray of flowers, but there are also many differences: 1. Classifiers are used in connection with all nouns, not just mass nouns. It is as if one is saying one piece [of] man instead of one man. 2. The meaning of many classifiers is weakened. In a way, one does not really say one piece [of] man, since the word corresponding to piece often has no clear meaning at all. 3. Unlike English cup, piece and spray, classifiers clearly constitute a separate word class, and most of them are never used as ordinary nouns. In Chinese, for instance, the noun for ’cup’ is be1izi, while the classifier for ’cup’ is be1i. 10 The correspondence between the count/mass noun distinction and the distinction between objects and substances is only approximate. In English, for instance, cloud is a count noun, although dividing a cloud in two does produce two clouds, and furniture is a mass noun, although cutting a piece of furniture in two does not produce two pieces of furniture. Note that many words have both count noun and mass noun usages, such as the mass noun oil (denoting the substance) vs. the count noun oil (pl. oils, denoting a type of oil).


20

PRONOUNS Words that typically ”stand for” a noun or a noun phrase are called pronouns. The most common pronouns are personal pronouns, like the following English forms:

3rd person 1st person 2nd person masculine feminine neuter

Subject form I he she Singular Object form me him her

it

Subject form we they Plural Object form us

you

them As far as we know, all languages have personal pronouns. Personal pronouns are always distinguished according to person (1st, 2nd and 3rd), though quite a few languages have two variants of the 1st person plural, an inclusive form (the ’we’ that includes ’you’) and an exclusive form (the ’we’ that excludes ’you’). Personal pronouns are also usually distinguished according to number (singular vs. plural; in some languages the dual and even the trial are given special forms), even in languages that do not distinguish singular and plural forms of nouns. The distinctions of case (subject form vs. object form) are much less common, though there are many languages with a much more fine-meshed case system than English. A related set of pronouns are reflexive pronouns, which are typically used to refer back to the subject of the clause of which they are a part. Compare the following two sentences: A. She killed herself. B. She killed her. In A, the reflexive pronoun herself must refer to the same person as the subject pronoun she, while in B, the personal pronoun her must refer to some other person than the subject pronoun she. English reflexive pronouns are differentiated on the basis of person, number and gender: myself, yourself, himself, herself, itself, ourselves, yourselves and themselves. Many other languages have only one reflexive pronoun (corresponding to English self), but often use it in conjunction with personal pronouns. The languages of the world differ from each other in the types of pronouns they possess. Consider the following two examples of such differences: 1. English has a number of possessive pronouns (my/mine, your/yours, his, her(s), its, our(s), their(s)), while other languages use personal pronouns plus possessive markers instead (cf. Japanese watashi no ’my/mine’ from watashi ’I’ and the possessive marker no). 2. English has four types of pronouns for which other languages, like Chinese, have only one. Depending on the context, many Chinese indefinite pronouns (she2i ’who(ever)’, she2nme ’what(ever)/which(ever)’, na3/ne3i ’which(ever)’) may be used


21

where English uses interrogative pronouns (who(m), what, which), negative pronouns (no, no one, none, nobody, nothing, neither), universal pronouns (all, every, each, everyone, everybody, everything) and indefinite pronouns (some, someone, somebody, something, either, any, anyone anybody, anything). It is not uncommon to call Japanese watashi no a possessive pronoun, it may be translated by English my/mine. It is evern more common to call Chinese she2i, she2nme, na3/ne3i interrogative pronouns, since they most commonly occur in contexts where English would use who(m), what or which. By doing so, however, we are in fact imposing an English terminology on languages for which they are not suitable. DYNAMIC VS. STATIVE VERBS Verbs may be dynamic or stative, depending on the kind of process they refer to: • A stative process is a sequence of uniform states. • A dynamic process is a sequence of heterogeneous states. If you make a movie of a dynamic process, you get a sequence of pictures where you can observe a gradual change from one picture to the next. If you make a movie of a stative process, you get a sequence of identical pictures. In FIG. 7, the three pictures in each row represent three states following each other in time, with the first state being the one on the left and the last state the one on the right. The earth going round is a typical dynamic process, while a house standing on the ground at the end of a small road is a typical stative process. A dynamic process A stative process 1st state 2nd state 3rd state FIGURE 7. A dynamic and a stative process Verbs like run, die and roll are dynamic, while verbs like have (in the sense of ’possess’), love and know are stative.

The difference between stative and dynamic processes plays an important role in the grammar of many languages. A number of grammatical constructions are fully acceptable with dynamic verbs like run, but less so with stative verbs like know: 1) Dynamic verbs have a progressive form, while stative verbs do not:11

a) John is running. b) *John is knowing the answer.

2) Verbs like force can be used with dynamic verbs, but not with stative verbs: a) I forced John to run. b) *I forced John to know the answer.

3) Dynamic verbs have imperative forms, while stative verbs do not. a) Run! b) *Know the answer!

11 Note how the McDonald’s chain makes the stative verb love into a dynamic verb by saying I’m lovin’ it.


22

4) Dynamic verbs may occur in the ”pseudo-cleft construction”; stative verbs may not:

a) What John did was run. b) *What John did was know the answer.

English has relatively few stative verbs, most of them denoting mental processes. Non-mental (as well as some mental) stative processes are usually expressed by adjectives (see below). In many languages across the world, however, there is no sharp distinction between verbs and adjectives. LEXICAL VS. AUXILIARY VERBS In English, a small subgroup of verbs called auxiliary verbs (or just auxiliaries) are clearly distinguished from other verbs (known as lexical verbs) in a number of ways: 1. Auxiliaries typically precede other verbs without the use of to (as in I hate to say this): a) You may go home. b) I have seen you. c) He is dancing. 2. They form questions by inversion rather than by means of to do (as in Do you study linguistics?): a) May I go home? b) Have you seen me? c) Is he dancing? 3. They are negated by the simple negator not, without the help of to do (as in He does not study linguistics): a) You may not go home. b) I have not seen you. c) He is not dancing. Many auxiliaries also have deviant patterns of inflection. While most languages have words resembling English auxiliaries in meaning and usage, the distinction between lexical verbs and auxiliary verbs is often much less clearly marked than in English. THE COPULA One verb with a very special position is the copula. A copula is used to link a subject and predicate, and it may imply a number of different semantic relations between the two, such as: 1. Identity: When the area behind the dam fills, it will be a lake. 2. Class membership: He is a bachelor. 3. Property: The house is blue. 4. Location: I am here. Some languages allow other semantic relations to be expressed by the copula, as when the Chinese copula shì is used to express a possessive relation:


23

Zha1ng xia1nsheng shì ge Rìbe3n ta4itai. ’Mr. Zhang is [one who has] a Japanese wife.’ A few languages have no copula at all. ADJECTIVES In English, adjectives also constitute a large and important word class. Semantically, adjectives typically refer to properties: red, hard, cold etc. Like entities and processes, properties may also be abstract, such as serious and brilliant. And like processes, properties are sometimes entified and referred to by nouns, e.g. happiness and brilliance. Syntactically, adjectives typically have two different functions, attributive and predicative. In English, attributive adjectives mostly precede the noun whose property they describe: red dust hard times cold water The opposite is the case in many other languages, such as French: jupe noire ’black skirt’ film interessant ’interesting film’ In English, predicative adjectives occur after verbs like be, become, turn and a few others: The dust became red Times are hard The water has turned cold Other languages, such as Hungarian, do not require a verb before a predicative adjective: Róbert öreg ’Robert [is] old’ As in many other languages, English adjectives have a comparative form (either happier or more happy) and a superlative form (either happiest or most happy). While nouns and verbs are more or less universal, there are many languages that do not have a significant word class corresponding to English adjectives. In Southeast Asian languages, for instance, most words corresponding to English adjectives are verbs. Red, hard, cold etc. are not only properties, but also states, and in this they resemble stative verbs, such as like or fear. ADVERBS The English words called adverbs constitute a very heterogeneous group. Semantically, they cover a number of different meanings, such as:


24

manner: slowly, badly degree: very, somewhat time: then, now place: here, everywhere cause: therefore, thus concession: nevertheless condition: otherwise Syntactically, they typically modify verbs (walk slowly), adjectives (really noisy), other adverbs (very badly), or whole clauses or sentences (He then left the country), but some words usually classified as adverbs are not modifiers at all (go home). Some types of adverb are closely related to other word classes. For instance, English manner adverbs are usually derived from adjectives, most often by adding the suffix -ly (as in slowly and badly), and they resemble adjectives in having comparative (more slowly) and superlative forms (most slowly). They are also semantically related, since adjectives may be seen as describing the properties of entities, while manner adverbs may be seen as describing the properties of processes. In English, there is also a strong connection between time and place adverbs and prepositions, the same word often being used as both adverb and preposition: Adverb: He went up and down. Preposition: He went up and down the road. In such cases, one and the same word simultaneously belongs to two different word classes. The class of adverbs, if it exists at all, is equally heterogeneous in other languages. ADPOSITIONS In their concrete meaning, English prepositions (like on, at, in, to, until, from, of, off, over, above, under, below, with, by, about, among, before, after, behind, through etc.) typically locate an entity in space or time: under the tree after midnight But prepositions are also often used in more abstract senses: under surveillance he is named after his grandfather The majority of prepositions may also be used as adverbs (see above). Prepositions are never inflected. When followed by pronouns, they are always followed by the oblique form (me, him, her, us, them) rather than the subject form (I, he, she, we, they). In other languages, such as Japanese, words corresponding to English prepositions are typically placed after the noun:


25

Tookyoo kara ’from Tokyo’ Kyooto made ’to Kyoto’ Such words are called postpositions. Even English arguably has three postpositions: ten years ago ten years hence ten miles away Both prepositions and postpositions may be referred to by the common term adposition. In some languages, adpositions do not constitute a separate word class. For instance, Chinese prepositions (like cóng ’from’ and za4i ’at, on, in’) are a subclass of verbs, while Chinese postpositions (like yîqia2n ’before’ and lîmia4n ’inside’ are a subclass of nouns. DETERMINERS In English, the class of determiners basically consists of the demonstratives this/these and that/those and the articles a(n) (indefinite) and the (definite). They typically stand at the very beginning or (in some languages) the very end of a noun phrase, making it more or less definite. Many languages have no articles, choosing either to leave the noun unmarked for definiteness or to use a numeral meaning ’one’ instead of the indefinite article and a demonstrative instead of the definite article. Quite a few languages have a larger number of demonstratives than English. For instance, both Japanese and Thai make the following three-fold distinction: ’this’ ’that’ ’that over there’ Japanese kore sore are Thai nîi nân nóon NUMERALS Virtually all languages have numerals. Some languages are reported to have very few of them, such as the Australian aborigine language Pitjantjatjara, which apparently only has the numerals ’one’, ’two’ and ’three’. Only the Amazonian language Pirahã is reported to have no numerals whatsoever. Like many languages, English makes a basic distinction between cardinal numbers (one, two, three etc.) and ordinal numbers (first, second, third etc.). CONJUNCTIONS Conjunctions are words that bind elements together within or beyond the sentence. Coordinative conjunctions like and, or, but are used to bind together words, phrases, clauses, sentences or stretches of speech larger than the sentence. The elements on each side of a coordinative conjunction are conceived as being of more or less equal weight, and often (though far from always) it is possible to change the order of the elements without changing the meaning of the whole expression: you and me vs. me and you.


26

Subordinative conjunctions (sometimes called complementizers) include one subordinate clause within a superordinate clause: If you touch me, I’ll shout. He told me that you would come. A girl that I knew appeared in the newspaper. In the first example, the subordinate if-clause has the same syntactic function within the superordinate clause as adverbs like tomorrow and otherwise, and such clauses are called adverbial clauses. In the second example, the subordinate that-clause has the same syntactic function within the superordinate clause as noun phrases like your story or many lies, and such clauses are called nominal clauses. In the third example, the subordinate that-clause has the same syntactic function as an attributive, and such clauses are called relative clauses. In English, many subordinative conjunctions are also used as prepositions and/or adverbs, such as before and after: He gave it to me before he left. (subordinative conjunction) He gave it to me before last summer. (preposition) He gave it to me before. (time adverb) But this is not necessarily the case in other languages. PARTICLES The term particle is used about any small, unstressed grammatical word that does not readily fall into any other word class. In Japanese grammar, for instance, the term is often used about phrase-final case markers (like wa for topic, ga for subject, wo for object and ni for indirect object) and sentence-final mood markers (like ka for question). INTERJECTIONS All languages appear to have a separate class of interjections, resembling English words like wow, psst, mhm and wham. The phonological make-up of interjections is often different from that of other words, and in any other word class, words like psst and mhm would have been impossible. Interjections may be divided into four basic subclasses: expressive (ouch, oh, wow, aha), directive (hush, psst, hey), phatic (mhm, yes, no, huh) and descriptive ideophones (wham, thud, bang). OPEN VS. CLOSED WORD CLASSES Of all the word classes examined above, some are open classes that readily admit new members, while others are closed and seldom admit new members. Nouns, verbs, adjectives, (some types of) adverbs and (in their own special way) interjections are the clearest examples of open word classes, while determiners, pronouns, classifiers, adpositions, conjunctions and auxiliaries are all closed sets that do not easily change.


27

CONTENT WORDS VS. FUNCTION WORDS To a large extent, the distinction between open and closed word classes coincides with the distinction between content words and function words. Content words are full words with a lexical meaning, denoting concrete or abstract entities, processes, properties etc. The meaning of function words, on the other hand, is more akin to the meaning of grammatical constructions. Usually, Often, the meaning expressed by function words in one language is expressed by grammatical constructions in another language. For instance, the meaning expressed by English prepositions (like in, at, from etc.) is often expressed by special forms of the noun called case forms in Finnish and Hungarian. Sometimes, even one and the same language has two alternative ways of expressing the same meaning, either with a function word or with a grammatical construction. For instance, more in more happy is equal in meaning to the comparative ending -er. The pronoun he is also very close in meaning to the verb ending -s (as in [he] runs), except that the latter contains no information about gender. 2.3.2 Word structure: inflection Morphology is the study of word structure, of how smaller meaningful units are combined into words, as in English sing-er-s. Morphology may be divided into two branches, inflection and word formation. The present subsection will be concerned with inflection. LEXEME AND WORD FORM We use the word word when talking about the vocabulary of the English language, as when we say Some people believe that the English language has more words than most other languages, probably because English dictionaries have so many pages. When talking about the word from the perspective of vocabularies and dictionaries, we shall introduce the term lexeme, reflecting the fact that linguists usually refer to the vocabulary of a language as its lexicon and that professional dictionary-makers are called lexicologists. We shall write lexemes with SMALL CAPITALS, usually followed by one or more subscript letters telling about the word class of the lexeme. For example, we have nouns like GIRLN and BOOKN, verbs like SINGV and WRITEV, and adjectives like GOODAdj and BADAdj.

In many languages, lexemes show up in different inflectional forms, which we shall refer to as word forms. For example, sing, sings, sang, sung, singing are the word forms of the lexeme SINGV and girl and girls are the word forms of the lexeme GIRLN. We may define lexeme and word form in the following way: A word form is a word with meaning and form. A lexeme is a family of word forms that are inflectionally related. We shall come back to the meaning of inflectionally related.

For several reasons, it is important to distinguish between lexemes and word forms. It would be very confusing to count word forms when you discuss the size of the vocabulary of a language. In Chinese, there is a lexeme FA;NGZIN ‘house’, which


28

has only one word form, fa;ngzi, since Chinese nouns do not have any inflection. In Turkish, there is a lexeme EVN ‘house’, which has 84 word forms. EVN has singular forms and plural forms, and among both the singular forms and the plural forms there are six different case forms, as shown in TABLE 7. Expression Meaning Expression Meaning ev ‘house; singular, absolutive’ evler ‘house; plural, absolutive‚ evi ‘house; singular, accusative’ evleri ‘house; plural, accusative’ evin ‘house; singular, genitive’ evlerin ‘house; plural, genitive’ eve ‘house; singular, dative’ evlere ‘house; plural, dative’ evde ‘house; singular, locative’ evlerde ‘house; plural, locative’ evden ‘house; singular, ablative’ evlerden ‘house; plural, ablative’

TABLE 7. Some word forms of Turkish EVN For example, evden means ‘from (the) house’ and evlere ‘to (the) houses’.

In addition, Turkish nouns may have six different possessive suffixes, telling who is the owner, and these suffixes can be combined with singular, plural, and all the case forms. A few illustrations are given in TABLE 8, where all the possessive suffixes are added to the absolutive forms (notice that possessive suffixes follows the plural suffix –ler), and in TABLE 9, where forms with the 1sg possessive suffix is combined with different case forms (notice that possessive suffixes precede the case suffixes). Expression Meaning Expression Meaning evim ‘house; 1sg, singular, absolutive’ evlerim ‘house; 1sg, plural, absolutive‚ evin ‘house; 2sg, singular, absolutive’ evlerin ‘house; 2sg, plural, absolutive’ evi ‘house; 3sg, singular, absolutive’ evleri ‘house; 3sg, plural, absolutive’ evimiz ‘house; 1pl, singular, absolutive’ evlerimiz ‘house; 1pl, plural, absolutive’ eviniz ‘house; 2pl, singular, absolutive’ evleriniz ‘house; 2pl, plural, absolutive’ evleri ‘house; 3pl, singular, absolutive’ evleri ‘house; 3pl, plural, absolutive’

TABLE 8. More word forms of Turkish EVN Expression Meaning Expression Meaning evim ‘house; 1sg, singular, absolutive’ evlerim ‘house; 1sg, plural, absolutive‚ evimi ‘house; 1sg, singular, accusative evlerimi ‘house; 1sg, plural, accusative evimin ‘house; 1sg, singular, genitive evlerimin ‘house; 1sg, plural, genitive evime ‘house; 1sg, singular, dative evlerime ‘house; 1sg, plural, dative evimde ‘house; 1sg, singular, locative evlerimde ‘house; 1sg, plural, locative evimden ‘house; 1sg, singular, ablative evlerimden ‘house; 1sg, plural, ablative

TABLE 9. Even more word forms of Turkish EVN ‘house’ For example, evim means ‘my house’ and evlerimiz ‘our houses’, while evlerimde means ‘in my houses’.

It would be meaningless to claim that for every Chinese noun there are 84 Turkish nouns, and that therefore Turkish has a much bigger vocabulary than Chinese. Vocabulary comparisons should be based upon lexemes, and the number of word forms in a lexeme does not influence the size of the vocabulary. MORPHEMES It is not difficult to discover that the Turkish word forms in TABLES 7–9 can be divided into smaller parts, each of which has its own meaning. These parts are called morphemes:


29

Morphemes are the smallest meaningful parts that words can be divided into. Here are the morphemes of the Turkish word form evlerimde ‘in my houses’: morpheme 1 morpheme 2 morpheme 3 morpheme 4 Expression ev ler im de Meaning ‘house’ ‘plural’ ‘1sg possessive’ ‘locative’

TABLE 10. The morphemes of the Turkish word form evlerimde ‘in my houses’ We find the morphemes of a word by comparing it to other words. We notice that ev occurs in all the words in the three tables, and the only constant meaning is ‘house’. We may also compare evlerimde ‘in my houses’ with ellerimde ‘in my hands’, where the morpheme el ‘hand’ can be identified. The element ler occurs in all and only the words having the meaning ‘plural’, im occur in all and only those having the meaning ‘my’ or ‘1sg possessive’, and de occurs in all and only the words with the meaning ‘locative’.

Now, take a look at some word forms of KÖYN ‘village’ in TABLE 11. We shall refer to a list of the word forms of a lexeme as a paradigm. This is therefore a partial paradigm for KÖYN. Expression Meaning Expression Meaning köy ‘village; singular, absolutive’ köyler ‘village; plural, absolutive‚ köyü ‘village; singular, accusative’ köyleri ‘village; plural, accusative’ köyün ‘village; singular, genitive’ köylerin ‘village; plural, genitive’ köye ‘village; singular, dative’ köylere ‘village; plural, dative’ köyde ‘village; singular, locative’ köylerde ‘village; plural, locative’ köyden ‘village; singular, ablative’ köylerden ‘village; plural, ablative’

TABLE 11. Some word forms of Turkish KÖYN ‘village’ Evi ‘house; singular, accusative’ contains the element i ‘accusative’, while köyü ‘vil-lage; singular, accusative’ contains the element ü ‘accusative’. In the same way, in ‘genitive’ is found in evin ‘house; singular, genitive’ and and ün ‘genitive’ is found in köyün ‘village; singular, genitive’. There is a system to it: we find i and in when the preceding vowel is i or e, and we find ü and ün when the preceding vowel is ü or ö. The plural forms köyleri ‘village; plural, accusative’ and köylerin ‘village; plural, genitive’ confirm this generalization. Cf. also the word forms gün ‘day; singular, absolutive’, günü ‘day; singular, accusative’, günün ‘day; singular, genitive’, disç ‘tooth; singular, absolutive’, disçi ‘tooth; singular, accusative’ and disçin ‘tooth; singular, genitive’ (the letter sç is pronounced [S], like in English dish [dIS]).

Few linguists would say that i ‘accusative’ and ü ‘accusative’ are two different morphemes, but rather two variants of the same morpheme. It is also very common to use the term allomorph instead of variant. The fact that morphemes have variants (or allomorphs) is rather unfortunate for our morpheme definition. One may claim that we find allomorphs when words are divided into meaningful parts, and that morphemes are not the smallest meaningful parts that words can be divided into, but rather families of allomorphs with the same meaning.


30

BASE AND OPERATION Let us continue our analysis of Turkish word forms, starting with evlerimde ‘house; 1sg, plural, locative’. We often think of word forms like this as having been formed by adding one morpheme to another, as illustrated in TABLE 12. Base ev 1st step: Add ler to the right of the base (= ev) ev+ler ⇒ evler 2nd step: Add im to the right of the base (= evler) evler + im ⇒ evlerim 3rd step: Add de to the right of the base (= evlerim) evlerim + de ⇒ evlerimde

TABLE 12. Bases and suffixes Each step consists of adding a morpheme to the right of a base; we shall come back to a definition of base immediately.

Adding a morpheme to the base is an example of applying a morphological operation to the base. Morphological operations can change the base in almost any logically possible way, by adding, modifying, and subtracting elements. A fourth type may also be established, which implies converting the base without any formal changes. We shall give examples below.

From our discussion it follows that a base is simply an element that a morphological operation can apply to. A base consisting of one morpheme only is referred to as a root. Therefore, the base ev in TABLE 12 is a root, while the bases evler and evlerim are not. ADDING ELEMENTS TO THE BASE A morpheme that is added to the base is called an affix, and there are affixes of three kinds: a suffix is added to the right of the base a prefix is added to the left of the base an infix is added into the base

Suffixes are the only affixes found in Turkish. A language with lots of prefixes is Swahili, as in the word forms m-toto ’child’ (singular) and wa-toto ’children’ (plural). Infixes are a much rarer phenomenon, but it is common inter alia in the Malayo-Polynesian language Tagalog of the northern Philippines, where, as shown in TABLE 14, the past tense morpheme is an infix in, which is inserted after the leftmost consonant of the base. present past ibigay ‘give; present’ ibinigay ‘give; past’ ipaglaba ‘wash (for); present’ ipinaglaba ‘wash (for); past’ ipambili ‘buy (with); present’ ipinambili ‘buy (with); past’

TABLE 14. Tagalog verbal word forms Infixes split morphemes into two parts, so that one part comes before and the other after the infix.


31

MODIFYING THE BASE In cases of modification, no special element is added to the base; instead, the operation changes one or more of the sounds in the base, as in English plurals like men (from man) and geese (from goose) and English past forms like spat (from spit) and wrote (from write). While such forms are exceptional in English, they are regular in Arabic. Arabic nouns are divided into two types, those that form the plural by adding an affix and those that form the plural by modification, which in this case implies changing the vowels, as shown in TABLE 15.

SINGULAR PLURAL qalb ‘heart; singular’ qulu\b ‘heart; plural’ kalb ‘dog; singular’ kila\b ‘dog; plural’ kita\b ‘book; singular’ kutub ‘book; plural’

TABLE 15. Some Arabic nouns SUBTRACTING ELEMENTS FROM THE BASE In the Nilo-Saharan language Murle, which is spoken in southern Sudan, the plural of nouns is formed by subtracting the last consonant from the singular form; cf. TABLE 16.

SINGULAR PLURAL nyoon ‘lamb; singular’ nyoo ‘lamb; plural’ wawoc ‘white heron; singular’ wawo ‘white heron; plural’ onyiit ‘rib; singular’ onyi ‘rib; plural’

TABLE 16. Some Murle nouns CONVERTING THE BASE Conversion means leaving the base formally unchanged (despite the meaning change), as in the English plural sheep (from sheep). Conversion creates synchretism, which is homonymy among the word forms of a paradigm. REDUPLICATING THE BASE Reduplication means repeating the base or part of it, as in Malay anak-anak ’children’ (from the stem anak ’child’). MORE COMMENTS ON OPERATIONS The plural forms of the Arabic noun kutub ’books’ cannot at all be divided into a morpheme meaning ‘book’ and a morpheme meaning ‘plural’, the way we can divide English books into book+s. Word forms formed by non-affixing operations are not easy to analyze within the morpheme model. 2.4.3 Word structure: derivation and compounding We will now turn our attention to the formation of new words. Consider the words player and football. Both consist of two morphemes, play-er and foot-ball. But there is a difference. In play-er, only the first morpheme may act as a word in its own right, whereas in foot-ball, both morphemes may act as words of their own. Both player and football are examples of word formation, but while player is an example of derivation, football is an example of compounding. Both player and football constitute inflectional bases, that is, the bases that inflectional operations (such as the adding of the plural suffix -s) are applied to:


32

players, footballs. Inflectional bases are also known as stems. In the study of word formation we analyze the morphological structure of stems.

If the stem cannot be divided into smaller meaningful parts (or if the stem cannot be analyzed as a base plus one or more operations), the stem is a root. Both play, foot and ball are roots, while player and football are the results of derivation and compounding, respectively. 7.3.2.1 DERIVATION The stem player can be divided into two morphemes, play and er. While play is both a stem and a root, however, er is neither, because it does not occur as the stem (inflection base) of any English lexeme. Instead er must be identified as a suffix added to verb stems to form stems of nouns designating actors. From the stem sing we get singer, and from the stem ride we get rider.

The suffix er is a derivational affix, that is, an affix added to a base to form a new stem. The stem player is derived from the stem play by suffixing er. In general, we may talk about derivational operations, which include affixing, modification, subtraction, and conversion.

The word form players (plural of player) has the morphological structure in FIGURE 14.

word form

stem

root derivational suffix

inflectional suffix

play er s

FIGURE 14. The morphological structure of the word form players COMPOUNDING The stem football can be divided into two morphemes foot and ball. Both of them are both stems and roots. Consequently, the stem football is a compound, that is, a stem formed from two or more other stems.

The structure of the plural word form footballs is presented in FIGURE 15.


33

word form

stem

stem stem

root root

inflectional suffix

foot ball s

FIGURE 15. The morphological structure of Turkish önsözde The two parts of the compound stem are also stems. In this case, they are also roots, though that does not have to be the case. 2.4 Sentences Syntax is the study of sentence structure, of how meaning is given form at the sentence level. In other words, syntax studies the systematic covariation between meaning and form in sentences. Most commonly, syntactic form has to do with word order. In addition, it is common and in many ways useful to group words into larger constituents and view each sentence as having a constituent structure. Finally, we shall have a brief look at other structural devices used for syntactic purposes, in particular agreement. 2.3.1 Word order In the English sentence (21), the three words vultures, eat, and hyaenas follow each other in a specific order:

(21) Vultures eat hyaenas. In this sentence, vultures is the first word, eat the second, and hyaenas the third. The most obvious aspect of word order has to do with precedence, with whether a given element precedes or follows some other element. In (21), for instance, vultures precedes eat, and hyaenas follows eat. In addition, word order also has to do with proximity. Compare the following expressions from English and Fula: 1. one big red ball 2. balloNre woÎeere mawnde woore ball red big one


34

The two expressions, which have the same meaning, consist of the same words, but the words come in exactly the opposite order. Even with regard to word order, however, there is a striking similarity between the two expressions. In both languages, the word meaning ‘red’ is closer to the noun meaning ‘ball’ than the word meaning ‘big’, and in both languages, the words meaning ‘red’ and ‘big’ are closer to the noun meaning ‘ball’ than the word meaning ‘one’. In neither language could these proximity relations be changed without creating ungrammatical or semantically odd noun phrases. In other words, the word order of the two expressions differ from each other with regard to precedence, but not with regard to proximity.

This is quite typical. Features of precedence vary a lot from language to language, while features of proximity are much less subject to variation. When comparing the word order of different languages, therefore, what is compared mostly has to do with precedence. 2.3.2 Constituent structure Take a look at the English sentences in (23).

(23) Some English sentences (i) Dogs bite postmen (ii) Angry dogs bite postmen (iii) Dogs bite angry postmen

It is easy to see that sentence (23i) may be divided in three, since it consists of three words. It is less obvious that (23ii) and (23iii) may also be divided in three, since each of them consists of four words. But in (23ii), angry and dogs clearly belong together, and in (23iii), angry and postmen also clearly belong together. Angry dogs and angry postmen are constituents of sentences (23ii) and (23iii), respectively.

Constituents like angry dogs and angry postmen are bound together semantically, but they are also syntactic units. In many respects, angry dogs and angry postmen work syntactically in the same way as the single words dogs and postmen. For instance, if they are placed in other positions in the sentence, they have to keep each other company, as is the case when some statements may be made into questions by reversing the order of the subject and the auxiliary verb:

(22) Some English questions (i) Dogs will bite postmen --> Will dogs bite postmen? (ii) Angry dogs will bite postmen --> Will angry dogs bite postmen?

With few exceptions, therefore, elements within the same constituent always occur in proximity to each other. In English, at least, angry cannot occur in positions further away from dogs, as illustrated in (24ii) and (24iii): (24) More English sentences

(i) Perhaps angry dogs bite postmen (ii) *Perhaps dogs angry bite postmen (iii) *Angry perhaps dogs bite postmen

Sentence (24ii) is possible to understand, but it is not a grammatical English sentence. The asterisk (*) marks the sentence as ungrammatical. In sentence (24iii), angry is


35

separated from dogs by the adverb perhaps, and the sentence is clearly ungrammatical. Angry is too far away from dogs, and it is probably impossible to decide who is angry—dogs or postmen. We would get a grammatical sentence by saying Perhaps dogs bite angry postmen, but this sentences would claim that it is the postmen that are angry, not the dogs. The sentence would be unsuccessful as an attempt to characterize the dogs as angry.

There is an asymmetry between angry and dogs, since angry is optional. It is possible to say Dogs bite postmen and Angry dogs bite postmen, but not *Angry bite postmen. When two words within a constituent are in an asymmetric relationship, the obligatory word is called a head and the optional word a modifier. Many linguists take one step further and claim that a head is the pivotal member of a specific type of syntactic unit usually referred to as a phrase: A phrase is a head plus its modifiers, if there are any. In the sentences (25), the phrases are placed between squared brackets, and heads are written in bold letters. Notice in particular that in (25i), dogs and postmen are regarded as phrases despite the fact that they are single words.

(25)English sentences again (i) [Dogs] bite [postmen] (ii) [Angry dogs] bite [frightened postmen] (iii) [The angry dogs] bite [the frightened postmen]

Phrases are classified into different types on the basis of the word class of their heads. If the head is a noun, we have a noun phrase—abbreviated NP—and if the head is an adjective, we have an adjective phrase—abbreviated ADJP. Adjectives can be modified by adverbs like very, slightly, extremely, to form adjectice phrases like [very angry] and [quite frightened]. In a more complete analysis, (26ii) and (26iii) should therefore be analyzed as in (26). A word is then analyzed as the head of a phrase if it can take modifiers, whether they are present or not.

(26) English sentences again (i) [[Angry] dogs] bite [[frightened] postmen] (ii) [The [angry] dogs] bite [the [frightened] postmen]

One consequence of the introduction of phrases into the syntactic analysis is the postulation of a hierarchical syntactic structure, implying the existence of linguistic units smaller than the sentence and bigger than the word. The term constituent is used as a name for a «sentence part» anywhere in the hierarchy. Sentence (26ii) may provisionally be analyzed as in FIG. 10, which, for obvious reasons, is called a syntactic tree.


36

sentence NP

NP

ADJP ADJP

The angry dogs bite the frightened postmen

FIGURE 10. Constituent structure presented in a syntactic tree First, sentence (27ii) is analyzed as having the three constituents, NP, bite, and NP. Secondly, the first NP is analyzed as having the three constituents, the, ADJP, and dogs, and the second NP is analyzed as having the three constituents the, ADJP, and postmen. Thirdly, the two ADJP’s have the single constituents angry and frightened, respectively.

A syntactic tree may also contain information about the word class of each word in the sentence:

sentence

NP

NP

ADJP ADJP determiner

adjective

noun

verb

determiner

adjective

noun

The

angry

dogs

bite

the

frightened

postmen

FIGURE 11. Syntactic tree including information on word classes

This makes it explicit which word is the head of each phrase, since the head of a noun phrase must be a noun, and the head of an adjective phrase must be an adjective.

Hierarchical syntactic structures have been postulated at least since the German psychologist Wilhelm Wundt (1832–1920) drew the first syntactic trees in his Völkerpsychologie ‘Ethnic Psychology’ (10 volumes, 1900–20). This way of viewing sentences has turned out to be very fruitful. However, there is in fact very little agreement about the exact constituent structure of even the most simple sentences, despite intense research through almost a whole century.

For instance, many linguists would claim that sentences containing a verb have a verb phrase (abbreviated VP), which, loosely speaking, consists of the verb plus all other constituents in the sentence apart from the subject. In our case, this would result in the following syntactic tree:


37

sentence

VP NP

NP

ADJP ADJP

The angry dogs bite the frightened postmen

FIGURE 10. Syntactic tree including a verb phrase Adherents of this analysis often define subject and object based on the position of noun phrases in a syntactic tree. Roughly, the object is defined as a noun phrase inside the VP, while the subject is defined as a noun phrase outside the VP. We shall return to alternative definitions of subject and object later in this chapter. 2.3.3 Participants In the subsection on sentence meaning above, we have introduced the notion of thematic roles like agent, patient, instrument, benefactive etc. In general, thematic roles are not marked directly in sentences, but through a fascinating interplay with syntactic functions, the most important of which are subject (S), object (O), and oblique (Obl). Take a look at the sentences in (13). (13) Some English sentences a) John opened the door with this key. b) This key opened the door. c) The door opened. The relationship between thematic roles and syntactic functions in the sentences in (13) is summed up in TABLE 4. SUBJECT VERB OBJECT OBLIQUE John opened the door with this key AGENT EVENT PATIENT INSTRUMENT This key opened the door INSTRUMENT EVENT PATIENT The door opened PATIENT EVENT TABLE 4. Thematic roles and syntactic functions In (13a), the subject is an agent, in (13b) a patient, and in (13c) an instrument. At first glance, this may give the impression that the subject varies so much semantically that it cannot be given a semantic definition, but this is not the case. SUBJECT AND OBJECT In fact, some important and central properties of the subject and object can be defined as follows:


38

The subject corresponds to the only participant or the participant closest to the energy source, while the object corresponds to the participant closest to the energy sink.12 In (13a), agent, instrument, and patient are mentioned. The agent is closest to the energy source, and is the subject. The patient is closest to the energy sink, and is the object. The instrument is expressed as an oblique, which contains a morpheme (the preposition with) expressing the thematic role. In (13b), instrument and patient are mentioned. The instrument is closest to the energy source, and is the subject. The patient is closest to the energy sink, and is the object. In (13c), patient is the only thematic role mentioned. It is only participant, and is the subject of the sentence.

Many languages accept sentences with two objects, and a few even sentences with three object. We shall number the objects from left to right. In the English sentence in (14), there is therefore a first object (O1) and a second object (O2). Here, we have introduced another thematic role, recipient, the receiver.

(14) First and second object SUBJECT VERB FIRST OBJECT SECOND OBJECT Mary gave John an apple AGENT EVENT RECIPIENT PATIENT Most linguists refer to O1 as indirect object and O2 as direct object, a terminology that cannot be extended to languages, like Fula, that allow three objects: (15) First, second, and third object in Fula SUBJECT VERB FIRST OBJECT SECOND OBJECT THIRD OBJECT Duudu hokkanii Buuba Muhammadu ceede Duudu hokk-an-ii Buuba Muhammadu ceed-e Duudu give-ben-pve Buuba Muhammadu money-pl agent event benefactive recipient patient ‘Duudu gave Muhammadu money for Buuba’ Abbreviations: ben = benefactive; pve = perfective. The last object is always the energy sink, and typically refers to an inanimate, non-sentient entity. The preceding objects are typically human (or at least sentient) beings that are not directly involved in the energetic interaction, which they rather relate to as experiencers. When there are three objects, the second is more involved in the energetic interaction than the first object. Typically, subject and object are realized as noun phrases, and languages can signal subject and object noun phrases by means of word order. Take a look at sentence (35), where the noun phrase [the hungry vultures] is the subject and [the dead jackal] the object.

(35) SUBJECT OBJECT [The hungry vultures] devoured [the dead jackal].

In English, the following generalizations are valid:

12 An exception is passive sentences like He was killed by his wife, in which the participant closest to the energy sink is a subject and not an object, while the participant closest to the energy source is oblique.


39

The noun phrase immediately preceding the verb is the subject. The noun phrase immediately following the verb is the object.

FIG. 12 shows the meaning and form of subject and object in English:

(27) subject (28) object meaning (29) ‘the only participant’ or

‘the participant closest to the energy source’ (30) ‘the participant closest to the

energy sink’ FORM (31) the noun phrase immediately preceding the verb (32) the noun phrase immediately

following the verb FIGURE 12. Subject and object as linguistic signs The meanings of subject and object are valid for most or all languages, while the form varies. Arabic is an example of a language where word order signals syntactic functions, but in a different way:

(36) An Arabic sentence SUBJECT OBJECT

T̆abaxat zawjatê at≥t≥aca\m ‘My wife cooked the food’ she.cooked my.wife the.food

In Arabic, the subject as well as the object follow the verb, but the subject is the noun phrase closest to the verb and the object follows the subject.13 OBLIQUE AND ADVERBIAL In English sentences, a participant may be referred to with an NP preceded by a preposition. A preposition + NP sequence is usually called a prepositional phrase, abbreviated PP. A participant consisting of a prepositional phrase has the syntactic function of oblique: (39) English sentences with PP oblique SUBJECT OBJECT OBLIQUE (a) [The old man] gave [a dime] [to [the little boy]]. (b) [The little boy] put [the dime] [in [his pocket]]. Now, compare this with sentence (40), where the PP contains an NP that does not refer to a participant, but to a setting (location). In such cases, the syntactic function of the PP is adverbial: (40) English sentence with PP adverbial SUBJECT OBJECT ADVERBIAL [The girl] met [her friend] [in [the train]]. Not surprisingly, words belonging to the word class of adverbs typically have adverbial function, as shown in (41). (41) An English sentence with an adverb as adverbial

SUBJECT ADVERBIAL [My friend] came [yesterday]

13 This generalization is only valid in sentences where the subject is expressed as a separate noun phrase.


40

How do we distinguish oblique from adverbial? As already pointed out, the oblique function is for participants, while the adverbial function is for settings. The oblique is an integrated part of the event, while the adverbial describes the scene where the event and its participants are «located», as illustrated by the two homonymous English sentences in (42). (42) Two English sentences: oblique versus adverbial

SUBJECT OBLIQUE ADVERBIAL (a) [The committee] decided [on [the boat]]. (b) [The committee] decided [on [the boat].

Sentence (42a) means that the committee reached a decision in favor of the boat, while (42b) means that the committee were on the boat when they reached a certain decision.

An important extra characteristic of the oblique is the collocational ties between the verb and the preposition. A collocation is a habitual juxtaposition of two or more words, like a verb and a preposition, and the preposition cannot be substituted by another preposition. In the case of the adverbial, there are in general no collocational ties, and the preposition can easily be substituted, as in The committe decided under the bridge. INFORMATION STRUCTURE Information structure may be defined in the following way:

Information structure is the way in which the words of a sentence are arranged so that a particular part of a message receives greatest attention.

Compare the two English sentences in (43). Notice in particular sentence (43b), which differs from (43a) by having two noun phrases before the verb. (43) Two English sentences

(a) [Bears] like [blueberries]. (b) [Blueberries] [bears] like.

Sentences never have more than one subject, and in English the subject NP precedes the verb immediately, with very few exceptions. Therefore, there is no doubt that [bears] is the subject in both sentences. In (43a), [blueberries] is definitely the object, since it is an NP immediately following the verb. What about [blueberries] in (43b)? It also has to be the object.

In English, as in many other languages, most phrases have their privileged position in the sentence, but may, in addition, occur in the beginning of the sentence, before the subject, as illustrated in (44). A phrase in this position plays the role of topic in addition to its ordinary syntactic function. For example, the money is object in (44a), object and topic in (44b).


41

(44) English sentences with a topic topic subject verb object oblique adverbial a John gave the money to Peter yesterday b The money John gave to Peter yesterday c To Peter John gave the money yesterday d Yesterday John gave the money to Peter Topic, which is grammatically marked by word order, may be semantically defined as the subject-matter of the sentence, that part of the sentence about which something is said. The rest of the sentence is then called the comment. (Some linguists use the terms theme and rheme instead.)

The reasons for putting a phrase in the topic position, for topicalizing it, are found in the wider linguistic context the sentence is used in. In conversation, topica-lization is often necessary to link a sentence in an appropriate way to what has already been said—cf. the constructed dialogues in (44) between Mary (M) and Paul (P). (44) Some dialogues 1st dialogue M: What happened to the money?

P: The money John gave to Peter yesterday. 2nd dialogue M: What did John give to Peter?

P: To Peter John gave the money (yesterday). 3rd dialogue M: What did you say happened yesterday?

P: Yesterday John gave the money to Peter. When Mary asks, What happened to the money?, she establishes the money as the subject-matter of the conversation, and Paul maintains the money as subject-matter by topicalizing it.

Peter’s answers are certainly not the only possible and appropriate ones, and instead of answering The money John gave to Peter he could have answered John gave it to Peter yesterday. But try to move the answers around a little, and the result is quite strange. If Mary asks, What did you say happened yesterday?, it would be very surprising if Paul answered, The money John gave to Peter yesterday, because it would destroy the cohesion of the text. CASE Case is variation in the form of pronouns, nouns or noun phrases to show their role in the structure of the sentence. In English, many personal pronouns distinguish between subjective case (as in I, he, she, we, they) and objective case (as in me, him, her, us, them). The subjective case is used in subject position, while the objective case is used in most other positions, notably object position and position within a prepositional phrase. Case languages typically distinguish between two or more of the following cases, with their typical syntactic and/or semantic features in parentheses:

nominative (subject) accusative (direct object) dative (indirect object) genitive (possessor) locative (at location) ablative (from location) allative (to location)


42

vocative () instrumental (instrument) absolutive ()

In many cases, however, cases do not make complete syntactic or semantic sense. Case may be expressed morphologically (by means of case affixes or special case forms) or syntactically (by means of function words, often referred to as case particles). 2.4.1 Events TENSE The grammatical category tense is related to the temporal setting of the ground, through a distinction between past, present, and future: past is a time preceding the temporal setting of the ground present is a time overlapping with the temporal setting of the ground future is a time following the temporal setting of the ground The time of the speech event, ‘now’, plays an important role in language. Most languages have a word meaning ‘now’, referring to the present; cf. the examples in TAB. 5. ‘Now’ is opposed to other times, ‘then’, away from the present and into the past or the future. – Closely related to ‘now’ are words referring to ‘today’, the day of the speech event. Following and preceding days are defined on the basis of their distance away from ‘today’—often, but not always, with different words for days in the past and days in the future. ASPECT The grammatical category aspect refers to the phase of an event, distinguishing between inchoative (beginning), progressive (continuous), perfective (completed) and others. Some languages, like Chinese, have only aspect and no tense, while other languages, like Russian, show a complex interplay between the two. MOOD The grammatical category mood refers to forms that are typically associated with specific speech acts, such as indicative (plain assertion: is), imperative (injunction: be!), and subjunctive (hypothesis: were), interrogative. VOICE The grammatical category voice usually distinguishes between active (He killed her) and passive (She was killed [by him]). 2.4.2 Agreement Agreement refers to a formal relationship between elements whereby a form of one word requires a corresponding form of another. In English, verbs in the present tense


43

agree with the subject in person (I go vs. he goes) and number (he goes vs. they go). Like word order and case forms, agreement helps to identify the roles or syntactic functions of each constituent in the sentence.

There are languages, among others Swahili and most other the Bantu languages of Africa, where the verb has not only subject agreement, but also object agreement (though only when the object is definite): (48) A Swahili sentence Adamu umekileta kisu. ‘Adamu has brought the knife.’ SUBJECT OBJECT Adamu u- me- ki- let -a kisu Adamu CL.1-PERFECT-CL.7-bring-INDICATIVE CL.7-knife In Swahili and other Bantu languages, nouns belong to different noun classes or genders, and the gender of a noun can in most cases be seen from a prefix, as in kisu ‘knife’, which has the ki prefix, marking noun class 7 (CL.7). The name Adamu has no prefix, but it belongs to class 1 (CL.1), the class of singular human beings. The verb form u-me-ki-let-a ‘s/he brought it’ has three prefixes. The first one, u, is the subject agreement prefix, showing that the subject belongs to class 1; the second one, me, is a tense/aspect prefix marking perfect, and the third one, ki, is the (definite) object agreement prefix, showing that the object belongs to class 7. The notion of noun class or gender is closely related to the notion of agreement: Noun classes or genders are classes of nouns that influence the shape of words that are in constructions with them. In French, for instance, nouns belong to one out of two genders, called masculine and feminine, and adjectives and determiners have to agree in gender with the noun they describe: un petit garçon ’a small boy’ une petite fille ’a small girl’ Le garçon est petit ’The boy is small’ La fille est petite ’The girl is small’ There are masculine and feminine forms of the indefinite article (un vs. une), the definite article (le vs. la) and the adjective petit/petite. While grammatical gender usually reflects biological sex, this is not always the case, and for the vast majority of nouns it makes little sense to speak of biological sex, and masculine and feminine are, in this context, purely grammatical terms.

Gender or noun class systems are very common in African languages. In grammars of languages in the Niger-Congo family, the term noun class is more common than gender, and the number of noun classes may reach twenty, as in Fula, where the semantic definitions of the classes are a fascinating study. In this language, the noun classes influence the shape of demonstratives, definite articles, adjectives, numerals and participles that are in constructions with a noun, and the demonstratives are used as class names. Consider TABLE 3, with nouns from the NDE class and the ngol class.


44

NDE CLASS NGOL CLASS Object yitere

hayre ‘eye’ ‘stone’

∫oggol mahol

‘rope’ ‘outer wall of a compound’

Place waalde ‘cattle camp, corral’ laawol ‘road’ Writing harfeere ‘letter of the alphabet’ diidol ‘line’ Process doÿÿere ‘fall; loss’ doÿÿol ‘falling; losing’ TABLE 3. Some Fula nouns The prototypical nouns of the NDE class designate relatively small, round, clearly delimited, three-dimensional objects like yitere ‘eye’, hayre ‘stone’, ∫ernde ‘heart’, and hoore ‘head’. This structure is imposed metaphorically on two-dimensional «objects», that is, places like waalde ‘cattle camp, corral’, wonorde ‘locality, abode’, and janngirde ‘school’ (literally, ‘reading-place’), and on the domain of writing; the smaller unities of writing belong to the NDE class, cf. harfeere ‘letter (of the alphabet)’ and hownde ‘suku\n (a small circle placed above a consonant letter in the Arabic alphabet to indicate the absence of a vowel)’. NDE class nouns also designate relatively short one-dimensional «objects» in the process domain, that is, clearly delimited episodes like doÿÿere ‘fall; loss’ and fuutere ‘fart’.

The prototypical NGOL noun, on the other hand, designates long, thin or narrow three-dimensional objects: ∫oggol ‘rope’, mahol ‘outer wall of a compound’, and meetaleewol ‘turban (that is, a long, narrow piece of cloth that is wound around the head)’. This long narrow structure is also found in NGOL nouns like laawol ‘road’; a road may be construed as a place that is narrow in one dimension and unlimited in the other dimension, and in the domain of writing, where NGOL class nouns designate long and narrow phenomena like diidol ‘line’ and binndol ‘writing’. Finally, in the process domain, NGOL class nouns designate events as phenomena without emphasis on the beginning and the end: doÿÿol ‘falling; losing’ and puutol ‘farting’. These examples show how processes, which are usually designated by verbs, may be entiified and talked about with nouns, and how these entified processes are structured in analogy with other «things». The difference between the delimited «NDE class process» doÿÿere ‘fall; loss’ and the unlimited «NGOL class process» doÿÿol ‘falling; losing’ is analogous to the difference between the delimited «NDE class place» waalde ‘cattle camp, corral’ and the unlimited (in one of its two dimensions) «NGOL class place» laawol ‘road’. NGOL class process nouns are mass nouns, while most other NGOL nouns are count nouns. You can count occurrences of a fuutere ‘fart’, while puutol ‘farting’ is uncountable. 2.5 Sounds Human languages may have existed for more than a hundred thousand years, and through most of this time languages have primarily been spoken. Writing was invented only five thousand years ago, and still only a few percentages of the 5,000–7,000 languages in the world are in daily use as written languages. Despite the existence of written languages—and sign languages (the “signed” languages of the deaf)—spoken languages have a privileged status that justifies treating it before written and signed languages.


45

We shall refer to the study of the sounds in language as phonology. The study of the sounds in language is usually divided into two parts, phonetics and phonemics, which can be defined in the following way:

PHONOLOGY The study of the sounds in language.

PHONETICS The study of the physiological, aerodynamic, and

acoustic characteristics of speech sounds.

PHONEMICS The study of how speech sounds are organized

into systems and utilized in languages. TABLE 19. The branches of phonology

In fact, most linguists these days refer to ‘the study of how sounds are organized into systems and utilized in language’ as phonology, while the term phonemics is not used at all, with the unfortunate result that they have no term for ‘the study of the sounds in language’ in general. We prefer to stick to the terminology introduced in TABLE 19. 2.5.1 Phonemes and allophones Throughout 2.5 we have written the pronunciation of words in a phonemic trancription, which is marked by slashes: /…/. The letters of a phonemic transcription represent phonemes, which we can define as follows: Phonemes are sounds that distinguish words (word forms) from each other. We may compare the English words pin /»pIn/ and pin /»tIn/, which are distinguished from each other phonologically by the contrast between the /p/ of the first word and the /t/ of the second. Consider also pin /»pIn/ and pan /»pQn/, which are distinguished by the contrast between /I/ versus /Q/, and finally pin /»pIn/ versus pit /»pIt/, which are distinguished by the contrast between /n/ and /t/. Pairs of words that are distinguished from each other by a contrast between two phonemes are called minimal pairs. As a general rule, a phoneme may vary in pronunciation from one position in the word to another. In English for example, /p/, /t/, and /k/ are aspirated (they are followed by a strong breath, like an h) in word-initial position, for example in pin /»pIn/, tin /»tIn/, and kin /»kIn/. On the other hand, they are unaspirated (they are not followed by a strong breath) when following /s/, like in spin /»spIn/, sting /»stIN/, and skin /»skIn/. The different pronunciation variants of a phoneme are called the allophones of that phoneme, and we have a special phonetic transcription to write them in. Such a transcription is written in square brackets, […]:

Orthography Phonemic transcription Phonetic transciption pin /»pIn/� [»pHIn]�tin /»tIn/� [»tHIn]�kin /»kIn/� [»kHIn]�spin /»spIn/� [»spIn]�sting /»stIN/� [»stIN]�skin /»skIn/� [»skIn]�

TABLE 25. Orthography, phonemic and phonetic transcription of six English words


46

Sound differences that are phonetic (or allophonic or non-distinctive) in one language may be phonemic (or distinctive) in another language. For English, the contrast between unaspirated and aspirated stops is a phonetic contrast, while there are many languages where the same contrast is phonemic, for example in many languages of India—cf. the Hindi words pa\l /pa˘l/ ‘take care of’ and ka\n /ka˘n/ ‘ear’ versus pha\l /pHa˘l/ ‘edge of knife’ and kha\n /kHa˘n/ ‘mine’. We have a tendency not to hear sound contrasts that are phonetic in our own language. The phonemes may be characterized as basic level terms of speech sounds, the highest level at which we can form a mental image of a sound concept. We have no difficulties in forming a mental image of an English /p/-sound, /t/-sound, or /k/-sound, but we are not able to form a mental image of a voiceless stop, which is the hyperonym term of which /p/, /t/, and /k/ are hyponyms. As we have observed earlier, terms above the basic level tend to have a more or less technical or scientific «flavor»— which is certainly true of ’voiceless stop’.

Even terms below the basic level tend to belong to the world of experts. And there is a level of hyponyms of the basic level phonemes, although the ordinary native speaker is seldom aware of its existence. 2.5.2 The speech organs Strictly speaking, human beings do not have «speech organs» the way we have «visual organs» (eyes) or «auditory organs» (ears). The speech organs, the totality of which we have also referred to as the vocal tract, is simply those parts of the mouth and the respiratory tract that are used to produce speech sounds. While seeing is the primary function of the eyes and hearing the primary function of the ears, the primary functions of the speech organs are breathing, eating, and drinking. The vocal tract may be construed as a “quartet” with the these members:

«The Vocal Tract Quartet» The lungs

& the windpipe The larynx The nasal cavity The pharynx

& the mouth The members of the “quartet”—except the lungs—are depicted in FIG. 16.

FIGURE 16. The speech organs (except the lungs)


47

THE LUNGS AND THE WINDPIPE By increasing and decreasing the volume of the lungs, air is caused to flow into and out of the lungs. In all languages some—often most or all—sounds are produced on the basis of air flowing out of the lungs and the windpipe (trachea), through the throat and out of the mouth and/or the nasal cavity. We can also use the tongue and parts of the throat to cause air to flow through shorter parts of the vocal tract. THE LARYNX The larynx is a box-like structure sitting on the top of the windpipe—cf. FIG. 17, where the larynx is seen from the front, with the structure known as the Adam’s apple in the middle.

Inside the larynx are the vocal cords, two flaps of muscle across the windpipe. The vocal cords can be held in different positions, affecting the airflow from the lungs in different ways. The opening between the vocal cords is the glottis, seen from above in PICT. 3. The vocal cords are the light folds surrounding most of the glottis.

FIGURE 17. The larynx PICTURE 3. The glottis and the vocal cords THE NASAL CAVITY The interior of the nose, the nasal cavity—cf. FIG. 16—is directly connected to the pharynx, so that air can flow from the pharynx, through the nasal cavity, and out. The passage between the nasal cavity and the pharynx—called the nasal port—can be opened and closed by lowering and raising the velum or soft palate, a muscular flap at the back of the roof of the mouth—cf. FIG. 16. THE MOUTH AND THE PHARYNX: THE ORO-PHARYNGEAL TRACT The mouth cavity and the pharynx (cf. FIG. 16), which is the tube starting right above the larynx (where the air-passage and the food-passage (oesophagus) divide on their way down) and going up to the back of the nose, constitute the oro-pharyngeal tract, which plays an important part in speech sound production.

Without comparison, the oro-pharyngeal tract is the quartet member with the heaviest and most challenging tasks. It is no coincidence that in many languages all over the world, language is referred to metonymically as «tongue» or «mouth». From a functional point of view, the mouth is the most important part of the oro-pharyngeal tract, and the tongue is the most important part of the mouth: without a tongue, no speech.


48

2.5.3 The basic components of speech production Speech production has two basic components: initiation and modulation. Without these components there is simply no speech. INITIATION Initiation is a movement of a speech organ to create an outgoing or ingoing flow of air through the vocal tract. The main initiation organs are the lungs, which work like a pair of bellows. When muscles are used to decrease the lung volume, the air pressure inside the lungs becomes higher than the air pressure outside the body, causing an outgoing airstream through the vocal tract. When muscles are used to increase the lung volume, the air pressure inside the lungs becomes lower than outside the body, causing an ingoing airstream through the vocal tract.

Airstreams can also be initiated by a piston-like movement of the larynx or the tongue: (1) Close the opening between the vocal cords and move the larynx up or down like a piston, or (2) close the opening between the back of the tongue and the back of the roof of the mouth and move the tongue backwards or forwards. The former type of initiation creates implosives and ejectives, while the latter type creates clicks. Implosives are found in many Asian and African languages, while clicks are only found in languages in Southern Africa. MODULATION Modulation is an organic posture or movement of a speech organ that changes the airstream in such a way that a sound is created. There are three types of modulation: phonation, nasality, and articulation. PHONATION Phonation is modulation in the glottis, which—as mentioned earlier—is the opening between the vocal cords. Two postures of the glottis are found in all languages: voiced posture and voiceless posture.

In the voiced posture, the glottis is quite narrow, so that the air-stream from the lungs causes the vocal cords to vibrate—cf. Fig. 18. Sounds produced with this type of phonation are called voiced sounds, like the sounds in the English words wall /»wÅ˘l/, Monday /»m√ndeI/, and busybody /»bIzibA˘di/.

In the voiceless posture, the glottis is wide open, and the vocal cords do not vibrate—cf. Fig. 19. Sounds produced with this phonation type are called voiceless sounds, like the consonants of the English words fat /»fQt/, post /»poUst/, and appetite /»QpItaIt/. The voiceless posture is also the posture of normal breathing.

FIGURE 18. The glottis in the voiced posture. FIGURE 19. The glottis in the voiceless posture.


49

Several other postures of the glottis are possible, including whisper, breathy voice, creaky voice, and glottal closure, which are treated in more advanced books. NASALITY As mentioned earlier, the nasal port may be opened and closed by lowering and raising the velum, the velum or soft palate, a muscular flap at the back of the roof of the mouth—cf. Fig. 6. Sounds produced with a lowered velum are called nasal or nasalized sounds, while those produced with a raised velum are called oral sounds. The term nasal is used when there is a complete closure somewhere in the mouth; if not, the term nasalized is used. Some linguists use nasal in both meanings. The consonants in the English word meaning /»mi˘nIN/ are nasal, while those in batik /b´»ti˘k/ are oral. Many languages all over the world distinguish between oral and nasalized vowels. In the French word chanter /SA)te/ ‘sing’, the first vowel is nasal (this is marked with the / )/ above the vowel sign) and the second is oral. Another example is Hindi /yaha”/ ‘here’. ARTICULATION Articulation is a modulation with a certain manner and place in the oro-pharyngeal tract.

Manner of articulation is type of stricture, while place of articulation is the location of the stricture in the oro-pharyngeal tract. For example, the initial consonant of the English word pen /»pen/ is produced with a complete closure (a manner) between the lips (a place), while the initial consonant of the English word sun /»s√n/ is produced with a narrow passage causing friction in the air-stream (a manner) between the tongue blade and the gums (a place).

As already mentioned, the articulation is the most complicated part of speech production. Manner of articulation is treated in more detail in paragraph 2.5.3 and place of articulation in paragraph 2.5.4. SUMMARY The division of labor between the members of “The Vocal Tract Quartet” is summed up in TABLE 20. Dark grey shows the most important function of a quartet member, lighter grey shows less important functions. initiation phonation nasality articulation lungs larynx nasal cavity oro-pharyngeal cavity TABLE 20. The functions of the “Vocal Tract Quartet” members. 2.5.4 Manner of articulation Speech sounds can be classified into several types on the basis of the manner of articulation. In a few cases, information about other modulations—nasality and place


50

of articulation—is also included in the classification. The following five types are most important: stops, nasals, fricatives, liquids, and vowels. STOPS Stops are defined by a complete closure in oro-pharyngeal tract and by a raised velum, preventing air from escaping through the nose. Stops are found in all languages. Voiced stop are a little less common than voiceless stops, and the former type is only found in languages with the latter type.

In English we find the voiceless stops /p t k/ and the voiced stops /b d g/, as illustrated by the words pet /»pet/, cat /»kQt/, bed /»bed/ and dog /»dA˘g/. NASALS Nasals are defined by a complete closure in the mouth and by a lowered velum, allowing air to escape through the nose. Nasals are found in the overwhelming majority of the languages of the world. Voiceless nasals are rare, and are only found in addition to the voiced ones. In English we find three nasals, /m n N/, which are all voiced, as illustrated by the word meaning /»mi˘nIN/. FRICATIVES Fricatives are defined by a narrow stricture in the oro-pharyngeal tract, causing audible friction as air passes by. Fricatives are common, but languages often have fewer fricatives than stops or nasals. Voiced fricatives are less common than voiceless ones, and are only found in addition to voiceless fricatives.

English has an exceptionally high number of fricatives, the four voiceless /f T s S/ and the four voiced /v D z Z/, as illustrated by the words thief /»Ti˘f/, fish /»fIS/, these /»Di˘z/, vase /»veIs/, vision /»vIZ´n/. LIQUIDS Liquids are a less homogeneous group than stops, nasals, and fricatives, but the term is a very useful common denominator for “l-sounds” and “r-sounds”. The typical liquid is produced with a constriction between the front part of the tongue (i.e., the tip or the blade; cf. 2.5.4) and the roof of the mouth; it is oral and voiced. The constriction is «lighter» than the one found in stops, nasals, and fricatives: either shorter than in stops/nasals or more open than in fricatives.

Most languages have at least one liquid. Korean and many Bantu languages have a single liquid that varies in pronunciation between [l] and [R]. [l] is a lateral, which is characterized by an air-stream passing the tongue on one or two sides through openings that are wide enough to avoid friction. [R] is tap, which is like a very short [d], that is, a momentary voiced stop: the tip of the tongue touches swiftly the roof of the mouth. Another common liquid is the trill [r], produced by letting the tip of the tongue vibrate against the roof of the mouth two or more times.


51

2.5.5 Place of articulation The initial consonants of the English words pan /»pQn/, tan /»tQn/, and can /»kQn/ are all voiceless stops, but they differ in place of articulation: The place of articulation of a speech sound is the place in the oro-pharyngeal tract with the strongest constriction during the production of the sound. To produce /p/, we make a complete closure between the lower lip and the upper lip; to produce /t/, we make a complete closure between the tip or the blade of the tongue and the alveolar ridge; to produce /k/, we make a complete closure between the back of the tongue and the velum.

At each place of articulation, there is a constriction between an active articulator and a passive articulator. The active articulators are the lower lip and the tongue, while the passive articulators are the upper lip, the upper teeth, the roof of the mouth, and the rear wall. Different languages utilize a different number of articulators, and the most important ones are presented in TABLE 21. ACTIVE ARTICULATORS PASSIVE ARTICULATORS labium labium apex dentes lamina alveoli dorsum postalveoli palatum velum uvular

TABLE 21. The active and passive articulators The active articulator labium (‘lip’ in Latin) is the lower lip; the apex (‘tip’ in Latin) is the tip of the tongue; the lamina (‘blade’ in Latin) is the blade of the tongue; while the dorsum (‘back’ in Latin) is the back of the tongue. Sometimes it is practical to use corona as a common denominator for the apex plus the lamina.

The passive articulator labium is the upper lip; the dentes (‘teeth’ in Latin) is the upper teeth; the alveoli (‘basins’ in Latin) is the alveolar ridge or the gums, minus the back half, which is called the postalveoli; the palatum and the velum are the hard palate and in FIG. 16, respectively. The lower part of the velum is the uvula. There are adjectives corresponding to the nouns in Table 21: labial, apical, laminal, coronal, dorsal; labial, dental, alveolar, postalveolar, palatal, velar, uvular.

Terms for active and passive articulators are combined to define places of articulation. For example, when the active articulator is apical and the passive articulator is alveolar, the place of articulation is apical-alveolar. Two terms are irregular; we use bilabial instead of «labial-labial» and labiodental instead of «labial-dental». Quite often it is practical to mention only the active or only the passive articulator. For example, alveolar is a common denominator for apical-alveolar and laminal-alveolar. For palatal, velar, and uvular sounds, it is superfluous to mention the active articulator, which in these cases is always the dorsum. In TABLE 22 we present some common sounds in the languages of the world, defined with parts of the terminology introduced so far.


52

bilabial labiodental

dental alveolar postalveolar

retroflex

palatal velar uvular

vc p� � � t� � ˇ� c� k� q�stop vcl b� � � d� � Í� Ô� g� G�

nasal vc m� � � n� � ˜� ¯� N� ≤�vc ∏� f� T� s� S� � C� x� X�fricative vcl B� v� D� z� Z� � � ƒ� “�

approx. vc w� V� � l� ®� Ò� j� � �tap vc � � � R� � � � � �flap vc � � � � � }� � � �trill vc � � � r� � � � � {�Abbreviations: vc = voiced; vcl = voiceless; approx. = approximant.

TABLE 22. Some common speech sounds. In TABLE 22, we have included a column with retroflex consonants, which strictly are sublaminal-prepalatal, that is, the active articulator is the area behind the apex, under the tongue, while the passive articulator is the front part of the palatum. However, the symbols for the retroflex consonants are also commonly used for apical-postalveolarisounds. Retroflex consonants are common in India, cf. Hindi d≥yu\t ≥ê /Íju˘ˇi˘/ ‘duty’. 2.5.6 Vowels While stops, nasals, fricatives, and liquids are consonants, we are now going to take a look at the vowels. The prototypical consonant has a relatively strong constriction in the oro-pharyngeal tract, while such a constriction is absent in vowels. Furthermore, the prototypical consonant belongs to the margin of a syllable (cf. the next paragraph), while the prototypical vowel belongs to the nucleus of a syllable.

Practically all languages distinguish at least three vowels, while the most common vowel inventory in the world is /i e A o u/, which is found for example in Swahili and Fula. We shall come back to a more detailed description of differences in vowel quality later. Vowels are traditionally analyzed with a terminology that differs from that used for consonants. Three main dimensions are used in this classification: • Horizontal tongue position: front / central / back • Vertical tongue position: close / close-mid / open-mid / open • Lip position: unrounded / rounded Before explaining how this terminology is used, we present TABLE 23, where some vowel symbols are placed in a chart in accordance with there classification. All cells could have been filled, but we have left out some less commonly used symbols. front central back unrounded rounded unrounded rounded unrounded rounded close i� y� � � µ� u�close-mid e� O� ´� � F� o�open-mid E� � � � � ç�open a� � � � A� �

TABLE 23. Some important vowel symbols


53

Let us start with i, which is front, close, and unrounded. Front & close means that the body of the tongue is pushed forward and upward, so that dorsum is as close to palatum as possible without creating audible friction. Unrounded means that the lips are spread, like in a smile. This is the vowel in English feel /»fi˘l/, where /˘/ tells that the vowel is long.

From i we move diagonally down to A , which is back, open, and unrounded. Back & open means that the body of the tongue is pulled backwards and downwords, so that the root of the tongue is as close to the back wall of the pharynx as possible without creating audible friction. This is the vowel in English father /»fA˘D´r/.

From A we move upwards to u , which is back, close, and rounded. Back & close means that the body of the tongue is pulled backwards and upwards, so that the back of the tongue is as close to the velum as possible without creating audible friction. Rounded means that the lips are rounded, with a posture resembling the letter O when you look at them in a mirror. This is the vowel in English fool /»fu˘l/.

Let us move to the fourth corner, to a , which is front, open, and unrounded. Front & open means that the body of the tongue is pushed forward and downward as much as possible. This vowel is found as the first element in the vowel in English price /»praIs/. This vowel is also found in French chat /Sa/ ‘cat’ and German da /»da˘/ ‘there’.

Between i and a you can move the dorsum up and down and stop at whatever height you like, but in this system two vertical tongue positions between close and open have been given a special status: close-mid and open-mid. These positions are obtained by moving the tongue in three equally long steps between the close and open: i – e – E – a. Here, e is the vowel of French bébé /bebe/ ‘baby’ and E is the vowel of French bête /bE˘t/ ‘animal, beast’.

In whatever position you have your tongue, the lips can vary between unrounded and rounded. For example, you can add rounding to i – e and get y – O, which are the vowels of French nu /ny/ ‘naked’ and nœud /nO/ ‘knot’; y is also found in Turkish üzüm /yzym/ ‘grapes’.

Let us then go the vowels u – o – ç , which are all back & rounded, but which differ by being close, close-mid, and open-mid, respectively. These are the vowels of French bout /bu/ ‘end’, beau /bo/ ‘beautiful’, and coteau /kçto/ ‘hillside’. If we remove the rounding of u – o , we get µ – F, which are the vowels of Vietnamese hu’ /hµ/ ‘spoiled’ and co’m /kFm/ ‘rice’; µ is also the vowel of Turkish kıptı /kµptµ/ ‘Gipsy’.

The central, close-mid, unrounded vowel ´ is the vowel of the first syllable of English again /´»geIn/. Central is the horizontal tongue position halfway between front and back. 2.5.7 Syllables Speakers of most languages are able to tell how many syllables there are in a certain linguistic unit. Speakers of English are generally able to tell that the words ape /»eIp/, baboon /bQ»bu˘n/, and gorilla /g´»rIl´/ have one, two, and three syllables, respectively.


54

Still, it is difficult to define the syllable in a simple way, and we shall simply present some characteristic features of the syllable.

The English word book /»bUk/ consists of one syllable, and may be said to have the structure in FIG. 20—that is, it has three parts, onset, nucleus, and coda. The onset and the coda constitute the margin of the syllable, that is, its non-nucleus parts. syllable onset nucleus coda /b� U � k/�

FIGURE 20. The syllable /»bUk/ The vowel is the only sound type that is allowed to occur in the nucleus in all languages, while only consonants are allowed to occur in the margin in all languages. Some languages accept more than one consonant in the onset and in the coda, but such congestions of sound are rather rare. European languages (and a few others) are exceptional here—cf. English words like strength /»streNT/, where the onset is /str/ and the coda is /NT/—but the languages of Europe constitute only about 3% of the languages of the world.

The only syllable type that is found in all languages is onset+nucleus, with one consonant in the onset and one vowel in the nucleus. There is a strong—but not exceptionless—tendency in the languages of the world for onsets to have a «richer» inventory of sounds than the coda. 2.5.8 Phonotactics As mentioned a few lines above, strength /»streNT/ is an acceptable word in English, but in most languages of the world it is an impossible word, for two reasons. First, because English has got sounds, for example /T/, that are quite rare; /T/ is found in Arabic and Swahili, but not in Hausa, Fula, Turkish, Modern Hebrew, Persian, Hindi, Indonesian, Chinese, Korean, and Japanese.

Secondly, the sound combinations in English are more complicated than in most languages outside Europe. The sound combination principles in a language are called the phonotactic principles of that language. They can be formulated on the basis of the word or the syllable. Consider the Japanese sentence in (49), where /. / marks syllable boundaries within words, that is, where one syllable ends and the next starts. Japanese onsets and codas never contain more than one consonant. (49) A Japanese sentence Zidoosya o motte Amerikazin wa arimasen ka? /dZi.do˘.Sa o mot.te a.me.Ri.ka.dZin wa a.Ri.ma.sen ka/ ‘Aren’t there any Americans who don’t own cars?’

chapter 2 from meaning to form - forsiden - universitetet i … · chapter 2: from meaning to form...

Documents